The Seven Magnificent

THIS PAGE ISBLANK

Copyright © 2005, New Age International (P) Ltd., PublishersPublished by New Age International (P) Ltd., Publishers

All rights reserved.No part of this ebook may be reproduced in any form, by photostat, microfilm,xerography, or any other means, or incorporated into any information retrievalsystem, electronic or mechanical, without the written permission of thepublisher. All inquiries should be emailed to [email protected]

ISBN (10) : 81-224-2324-8ISBN (13) : 978-81-224-2324-2

PUBLISHING FOR ONE WORLD

NEW AGE INTERNATIONAL (P) LIMITED, PUBLISHERS4835/24, Ansari Road, Daryaganj, New Delhi - 110002Visit us at www.newagepublishers.com

Dedicated to the cause of prosperityreleased during 2004

the 111th birth anniversary yearof Prof. P.C. Mahalonobisthe Founder Director of

Indian Statistical Institutethe temple of my learning

the centenary year of Dr. J.M. Juranthe world quality Guru

the 84th year Dr. C.R. Raothe best living statistician

and my esteemed guru.

THIS PAGE ISBLANK

FOREWORD

Indian industry has made tremendous progress in the last three to four decades. It hasestablished manufacturing capacity in wide ranging products and technologies including thosein high-tech areas like space, computers and biotech. In terms of quantitative productionIndia ranks among the top few countries. However, on the quality aspect the picture isentirely different. In terms of exports which is true indicator of quality and costcompetitiveness, India ranks even lower than small countries in South East Asia whichstarted industrialization much later than India. This is in spite of the solid advantages of vastnatural resources, highly competent managerial and technical manpower and low cost oflabor.

Main reasons for this seeming paradox is that managements of Indian industries pay onlylip service to quality management under mistaken belief that emphasis on quality would addto costs. This myth has been shattered by a number of controlled studies in Europe andJapan. The only way Indian industry can take its rightful place is through systematicallyplanned quality improvement projects which can significantly improve quality conformanceand also lead to cost reduction. It is here that the well-tried seven statistical tools can be usedto make tangible improvements in quality and cost saving through elimination of wastes of alltypes.

One reason why these statistical tools are not fully utilized even in professionallymanaged companies is that these are perceived to be difficult to understand and applybecause of advanced mathematics and statistics. Prof Nankana has done commendable serviceto industry by explaining these tools in simple language which even persons with high schoolbackground can follow. He has not stopped at explaining the concepts of these tools, but hasgone much further by providing the methodology of their deployment. To convince theskeptics in industry, he has backed up application methodology with real life case studiesbased on his vast experience of consultancy with industries in different sectors.

I have had long association with Prof Nankana going back to our college days. During mytenure as Director General of Defence Quality Assurance and Bureau of Indian Standards, Ihad the opportunity to work closely with Prof Nankana on policy issues for quality promotionand development of National Standards on Quality Management. He always had down toearth shop floor approach and put across his views with passion, which cannot fail to impress.We can see the same approach in this book which is bound to motivate even the die hardconservatives to try these tools for improving the quality of products and processes.

This book has filled a definite void in Indian quality management literature and I amsure business managers and quality professionals will take advantage of it to put quality inIndia in higher orbit.

Lieut General H. LalDirector General FICCI Quality Forum and

Chairman National Quality Campaign CommitteeQuality Council of India

APTITUDE TEST TO ASSESS ASSIMILATION 7

d:\data\newage~1\Magni\Ma-co.pm5/Third proof/4-12-04

PREFACE

It is being recognized more than ever before that the Quality is the Sharpest Weapon tosurvive, compete and prosper. It calls for eternal endeavour of alternating sequences ofactivities for control and break through. This virtually implies imbibing Quality Culture by alland sundry in the organization.

Efforts need to be focused on analysis of true, adequate and relevant data to be efficientlyeffective. This in turn requires scientific bent of mind, may we call it a Statistical Culture.

Statistics is a science in search of truth. It is a science that serves all other sciences and ismaster of none. It is a key technology consisting of preparing a problem bank, prioritisingthese, collecting relevant right amount of data, analyzing these and making recommendations,implementing these, stabilizing control and repeating the cycle to achieve the goal of being aleader, through continuous improvement both evolutionary and revolutionary.

It is most unfortunate that statistics is, by and large, considered a very dull, dry andcomplex subject, so much so that it is shunned. An humble attempt has been made in thissmall volume of 141 pages to introduce seven tools that are simple, elementary, easy to learnand practice quickly, so diverse in application and yet very highly cost effective with very vastpotential for abundant returns. These are all fascinating. The emphasis has been placed ontheir concepts and applications through illustrative case examples with a pious hope that itwill serve the need of developing self sufficiency in the use of statistical methods forcontinuous improvement, at all levels in all fields of activities. An attempt has been made tobring home the truth that the fruits of applying statistics are the sweetest. There is no escape.Therefore endear it, the sooner the better, for the prosperity of the society. Hopefully, it shouldgenerate interest beyond these tools.

Magnificent seven refer to the seven tools namely Cause and Effect Diagram, Check Sheet,Pareto Analysis, Stratification, Scatter Diagram, Histogram and Run Chart. These are treatedin chapters 2 to 8. These are preceded by a chapter describing the associated terminologies,concepts & economic significance and succeeded by chapters on a composite case study usingall the seven tools, Organising for quality control and tips to imbibe Quality As a way of lifethrough its ABC to habitually steer on the path of Continuous Improvement. The last chapterprovides for Aptitude Test to assess Assimilation and the gap that needs to be bridged. TheseElementary Basic Tools applied creatively embrace bulk of the problems, often faced in day today work.

This booklet is the cumulative result of the dedicated and committed inputs received fromthe respected parents, teachers, literature, colleagues and the students in the class room and

8 THE SEVEN MAGNIFICENT


at work places consisting of several customer organizations, as also all members of my family.A select bibliography of authors and books is given at the end. I record my grateful thanks toall of them including human resource of associate organizations.

I request the readers to contribute to further enhancing its value by providing usefulfeedback after reading the contents and putting these into practice and earn my gratitude aswell as that of the readers of subsequent revised versions.

A N NANKANA

Patron, Quality Improvement Mission, New Delhi

APTITUDE TEST TO ASSESS ASSIMILATION 9


CONTENTS

Foreword vii

Preface ix

1. Fundamentals of Quality 1

2. Cause and Effect Diagram 10

3. Check Sheet 17

4. Pareto Analysis 24

5. Stratification 33

6. Scatter Diagram 45

7. Histogram 61

8. Run Chart 75

9. Composite Case Study 99

10. Organising for Quality 112

11. Imbibing Quality, As a Way of Life, through its A B C 119

12. Aptitude Test 133

List of Case Studies 142

Bibliography 149

Other Forewords 155

THIS PAGE ISBLANK

FUNDAMENTALS OF QUALITY 1

d:\data\newage~1\Magni\Magni-1.pm5/Third proof/4-12-04

1FUNDAMENTALS OF QUALITY

1.1 WHAT IS QUALITY?

Quality has been universally acclaimed as fitness for use, signifying the supremacy of the userin judging its adequacy. For instance it is for the user to decide how good a product is. Onlythe wearer knows where the shoe pinches. Likewise the user can tell how good is a pencil, amixer grinder or a gear. Further, whether he is satisfied with the electric supply, its voltageincluding its stability and continuity.

● It refers to all types and sizes of products and services and in fact any activity. Producthence forth shall include service also.

● It is a composite result of a satisfactory design and its conformance.● Conformance to specification(s) of product parameter(s) is (are) indirect approximate

means of forecasting adequacy of product to satisfy the intended use. Productparameters include chemical, durability, mechanical, physical and allied properties.Like appearance, availability, brightness, colour (shade), composition, cost, diameter,ease of maintenance and replacement, fastness of colour, hardness, odour, purity,safety, timely delivery at desired place. This list is not exhaustive.

● Quality therefore encompasses all features like :Appearance : that influences first sale.Functional : that causes repeat sales andReliability : including availability and maintainability that sustains the market to ensure

growth for survival. Lack of it is fatal sooner than later. Availability implies that the gadgetworks when operated or switched on while maintainability implies that repair, as and whenit becomes necessary, is easy, economical and quick.

● Formal definition by International Organisation for Standardisation ( ISO ) read as:

Totality of features and characteristics of a product or service that bear on its ability tosatisfy stated and Implied needs.

It now reads as

Degree to which a set of inherent characteristics fulfils requirements.

The preceding discussions only vindicate the above statement.

2 THE SEVEN MAGNIFICENT SIMPLE, QUICK AND COST EFFECTIVE...


1.2 WHY QUALITY?

Consider mixer/grinder, for instance. A housewife shall be tempted to buy it, if it looks nice.She buys one, takes it home and operates it. It may not function as expected. She tells othersabout her experience. This message spreads like a wild fire and puts brake on its further sale.On the contrary, if she is satisfied, sharing her experience with others, will promote repeatsales. Further, as and when the service becomes necessary, it is reliable and effective life cyclecost is competitive, the organization not only sustains the market but is also able to ensurehealthy growth rate. This generates more employment opportunities and hence prosperity forthe society.

Quality indeed makes a difference between success and failure. Thus quality is necessaryfor survival, though survival is not mandatory.

1.3 WHAT ARE THE MEASURES OF QUALITY?

Some of the key indices of quality achievement are:

● Degree of satisfaction as reported by the user, Market complaints and Returns.● Productivity Indices, namely Ratio of Conforming Output to each Input.

Inputs include both direct and indirect. For example, energy, equipment, humanresource, raw materials, and the invisible yet instantly perishable time. Each indexneeds to be monitored appropriately.

● Ratio of Ideal versus Actual costs of production .The Ideal cost of production is estimated under the assumption of Zero Wastage ofinputs from conception of product to its fruition by doing every thing Right First Time.

● Measurement of Quality is the Price of Non-conformance.Account for every thing that need not have been done or would have been avoided ifevery thing were done the right way in the first instance and consider that as the Priceof Non-conformance.Therefore, negatively, it is measured by the people with vision, as the loss that iscaused to the society by the lack of quality or imperfection in the product delivered orservice rendered.Imagine a Power House generating power lower than the designed capacity. The socialloss is not just the loss of revenue to the producer of power, it also includes the loss ofopportunity to provide better service to the existing users, and possibly meeting extrademand domestic or industrial. The latter adds to more employment potential providingnecessary thrust to enhance Gross National Product and welfare of the society.

● Quality may also be simply expressed in terms of percent non-conforming units ornonconformities. Further if the criticality of various types of non-conformities is notuniform viz; the likely potential loss or damage caused is not equal or varyconsiderably, then an appropriate weighted demerit score is used as an index to signifythe level of quality achieved. Inspection error, classifying conforming as non-conformingand vice versa, should not exceed one tenth of Acceptable Quality Level. Ideally thequality is zero non-conformity.

● Quality of an item with respect to any one measurable parameter may be assessed byits proximity to standard or target namely specified mean (middle value of themaximum and minimum specified acceptable limits called tolerance limits).



● Quality of a lot is best defined by the pattern of proximity of deviations of theobservations from the target designated as zero. It should conform to the normalpattern or any other expected pattern depending upon the parameter being assessedand not violate the specified tolerance band. The process mean and standard deviation(root mean square deviation) together describe the quality. The smaller the variance(square of standard deviation) the better the quality. Ideally it ought to be zero. InsteadDr. G Taguchi�s loss function, quantified as proportional to square of deviation,expresses quality in terms of monetary loss. This successfully highlights qualityimperfections even though the product may be conforming to the specified norms. Thismight have been the single most important factor in focusing the attention of the topmanagement on quality as hitting the target rather than satisfying the specification.This itch probably made Japan the world leader. It is obvious that loss arrived atthrough loss function is under estimate in comparison to the visualized social loss.

● Consider the game of hitting the bull�s eye among four players A, B, C, and D.Compare the following four emerging situations. See Figure F1.1.(A) is hitting away from the target and the hits are spread over a wide area.(B) is hitting the target but the hits are spread over wide area like that in (A).(C) is off the mark like (A) but the spread is over narrower area. He has bettercapability than (A). His failures are due to being off the mark. His performance will bebetter than that of (B) if only he can correct his bias, which is considered easier thanimproving one�s intrinsic capability.(D) is hitting the target and simultaneously in closer range too like (C). Theperformance Quality of hitting the target is best in this case.

FIGURE F1.1 Comparison of performance of four players hitting the bull’s eye.

In defence parlance, if one fails to hit the target , it survives to blow the fatal hit inreturn.

Thus quality is measured positively by the value, the user attaches to what in hisperception he is receiving in return for the rupee (s) spent by him or negatively by the loss thatlack of quality is likely to impart to the society.

CD

AB



It may also be expressed by percent non-conforming units or non-conformities or byappropriate demerit score or by the proximity of the value of the parameter of interest to thetarget together with its spread around it.

1.4 DOES BETTER QUALITY COST MORE?

It is often believed that superior quality of design costs more because of necessary costlierinputs. This is not necessarily true. Let us consider a practical example of two well knownbrands of scooters in India, Vespa (Bajaj) and Lambretta (Vijay super). Both these aredesigned to cater to the specific travel needs of particular economic group or market segment.Yet, one of these decisively established its overall superiority in design at lower cost. Moreexamples are possible from housing designs to satisfy the needs of the customer in all respectsat lower cost. Thus it is possible to have better design at lower cost.

Like wise, it is a common belief that production of conforming product entails stoppage ofprocess for correction of errors that reduces production and productivity and enhances costper unit of conforming product. Thus suggesting that quality of conformance also costs more.This too, is false. Consider the consequences of having produced a non-conformity and hencea non-conforming item. It has to be either reworked or scrapped. The former requires re-handling and re-inspection, increases inventory of in-process goods and re-processing. Re-processed goods are often not as good as the originals. If, however a non-conforming unit slipsthe inspection system and reaches the customer, it adversely affects the market share,generates fire-fighting and might lead to even struggle for survival besides involving extracosts to compensate the customer including free replacement and fulfill the legal andstatutory obligations, if any. It also places additional burden of enquiry and investigations todetect and correct the source of error and to put the appropriate fool-proof system in place,including training, to avoid its recurrence. The latter alternative of scrapping the item implieswastage of all the resources consumed and continued generation of losses till the snag causingthe problem is removed. Together the total cost and associated repercussions constituteHerculean prohibitive task. Non-conformities are not free. It costs to produce these. It costsextra to eliminate these and their associated side effects. The cost of prevention ofnonconformities and thereby nonconforming units should be compared with total losses, thatare likely to be caused to the society by these imperfections, that shall vanish if these wereavoided. Hence, quality of conformance always costs less, lack of it more.

Thus quality is always an economically viable alternative. Quality control aims at andachieves improved quality at reduced costs. This way we have best of both worlds. Utility ofproduct or service per rupee is the real index of quality, the higher the better.

1.5 WHAT ARE QUALITY PROBLEMS?

The problems are in abundance. One needs to create a problem bank, prioritise these,organize teams with proper facilitation and empowerment to resolve these as per plannedschedule.

The quality problems fall in two distinct categories. These are sporadic and chronic.Let us look at Figure F1.2. This shows sequence of production on X- axis and quality

index on Y- axis, say percent non-conformities, the lower the better. The non-conformities in



fifth batch are too low and in thirteenth batch too high to be ignored. Such occurrences hereand there arise from what are termed SPORADIC problems. These have roots in lack ofprocess control. The process is suspended and reasons for the abrupt change identified. Theseneed to be inculcated in the former case, if economical and avoided in the latter case tobenefit in either situation. These measures are parts of process control. Its organizationinclude the following steps:

Assess Process Capability (Variation due to Chance Causes only)Fix Optimal TargetChoose Appropriate ChartDetect Abnormal DeviationInvestigate the CauseRestore the statusIdentify Uncontrollable Assignable Factors. Monitor these and Manipulate, such among

the Rest Appropriately that counter the harm-full effect adequately. Reference is invited to IS:397 parts 0, 1, 2, 3 and 4.

FIGURE F1.2 Illustrating the existence of sporadic problem.

Now consider another situation where a competitor organizes special studies and succeedsto improve product design that enhances product worth or process design that enhancesdegree of conformance and or reduces cost. See Figure F1.3. He can market his productcompetitively and pose challenge to others. Further his gain from this improvement is eternalfor all future time to come. These situations are best described as chronic problems andapproach to achieve this better level of performance is called break through. The steps toexecute this programme include the following:

Developing positive change in the attitudeLaying down priorityDeveloping and executing relevant training programmesConstituting Steering and Diagnostic ArmsBringing in desired Cultural changeOrganising Switchover to put the improved system in place

2

4

6

8

10

12

14

02 4 6 8 10 12 14 16 20 23 26

Sporadic

Batch No.

X

Y

Sporadic

Non

conf

orm

ities



The cycle of control (of process) and break through should form a permanent feature ofevery organization, since quality is a journey and not a destination.

One needs to plans one�s path to resolve problems in time bound programme.Any perception, at this stage that all is well and, that there is no problem, signals the

start of the fatal journey. In the absence of any apparent problem, the problem ofbreakthrough always exists. There is always a challenge to do better and occupy the space atthe top. In fact when there is no crisis that warrants fire fighting and all is peaceful, it is thebest time to attempt improvement of breakthrough type, with a cool mind free of any tensionwhatsoever.

The problems listed in problem bank are prioritized. Once the priority list has been made,on the basis of the harm that it causes; the steps, shown in Figure F1.4, are proposed toresolve the problem on hand.

It is imperative to choose more than one appropriate indices to measure or assess theeffectiveness of improvement. Often a single index can be misleading. The improvement maybe taking place at the cost of some other adverse effect with overall loss increasing. Forexample, the inventory might increase disproportionately to the benefit accrued fromenhanced availability of the material. One may begin with a couple of indices.

Having decided on the indices, one needs to go in for planning the generation of data thatwill help valid calculation of the indices and provide link with the causative factors.

All the tools, that form the subject matter of this booklet, learnt should be innovativelyapplied to analyse the data gathered to derive maximum information and chalk out acomprehensive plan of action for improvement.

Take action(s) and review the improvement attained against the projected one�s.

02

4

6

8

10

12

14

16

18

20

22

24

26

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27

Break through

X

Y

Month

Pro

duct

ion

cost

Rs.

per

unit

FIGURE F1.3 Illustration of chronic problem fit for breakthrough.



The gap constitutes the basis for next iteration to repeat the cycle of activities forresolution of the problem for further improvement.

May it need be reminded, that, any glaring deficiency observed that can cause animperfection or contribute to the problem on hand, even though unintended, need to have beensatisfactorily attended to, before resorting to any structured approach to resolving the problem.

The culture of facing a problem as it arises is synonym with fire fighting. One successfullyresolves the problem perhaps very speedily too, but then the organization and the society paysa very heavy price. The fire fighter might earn his promotion too! On the contrary, a systemof prevention needs to be practiced as a way of life. This culture shall avoid the menace offirefighting. The system consists of anticipating the problem, developing a fool proof system toforestall it and to put the system in place.

1.6 WHAT IS THE ROLE OF STATISTICS?

Before attempting answer to this question, it is only fair to understand, what is statistics?It is a science in search of truth and truth alone, nothing else but truth. It is a great pity

that yet one of the quotes is; lie, damn lie, white lie and statistics. Statistics never tells lies.Some statisticians might do in the circumstances they are placed. Then telling lies is not theirmonopoly. It is not being said in their defence, but again a statement of facts. No doubt, theprofession demands highest order of integrity.

Statistics is a science that serves all other sciences and is master of none. It delves into thevoid, to know the unknown and traverses from uncertainty to certainty. The un-intended risksarising from sampling and non-sampling errors are contained within acceptable norms withdue consideration of long term economic impact.

The word statistics like several others has multiple meanings. Commonly it means anydata. For example, Agricultural Statistics, Business Statistics, Commercial Statistics, DefenceStatistics, Economic Statistics, Educational Statistics, Health Statistics, Intelligence Statistics,

I D A A R A (Organisation)

Indices

Data

Analysis

Action

Review

Again

FIGURE F1.4 Steps to resolve the problem.(The urdu word IDAARA means an organization)



Population Statistics, Railway Statistics, Transportation Statistics, Vital Statistics and so onand on.

Specifically the word statistics means an appropriate function of observations made on asample of units selected suitably to represent the population to estimate the correspondingpopulation parameter. Based on sample size and its method of selection, its confidence levelcan always be determined. Conversely, sample size can be determined to provide an estimateof desired accuracy and confidence.

Statistics has also been defined as a key technology. It consists of Formulating theProblem Precisely, Gathering Adequate Relevant Representative Data, Aptly Analysing,Validly Concluding, Making Confirmatory Trials, Making Recommendations Based on theCumulative Findings, Implementing these to Reap the Expected Gains, Assessing the Gap(s),Repeating the Cycle to Bridge the Gap(s) and Continue the Chain of improvements to Reachthe Evasive Goal of perfection�may be Zero non-conformity, Zero Wastage, Zero Deviationfrom the Target and the like. No wonder ISO : 9000 family of standards on Quality Systemsmakes use of Statistical Methods Obligatory.

Statistics provides indispensable scientific tools to solve problems of quality controlincluding break through to sustain continuous improvement. The Basic tools of StatisticalQuality Control include Cause and Effect Diagram, Check Sheet, Pareto Analysis,Stratification, Scatter Diagram, Histogram and Run Chart. These are Simple, Easy and Quickto learn and practice fruitfully. Reference is invited to IS : 15431. There are a host of othertechniques to develop optimal solutions to problems encountered in almost all variety ofsituations.

Statistical methods are available to assess inspection inaccuracies arising from samplingand non-sampling errors. For the cumulative effect to be harmless, the error ought to be lessthan one tenth of acceptable quality levels and in no case in excess of one sixth. Generallythese errors are found to be on the higher side. Statistical studies have helped to, reducethese to acceptable norms, standardize, and sustain these. These studies fulfill the obligationsunder the title Repeatability and Reproducibility Errors pertaining to ISO : 9000 family. Thisactivity should precede planning for Process Control System and putting it in place. Statisticalaids for process control have been mentioned in preceding section (1.5)

Statistical approach of experimentation for determination of optimal product and processdesigns have been exploited massively by advanced countries and in a limited way bydeveloping one�s. These help in getting valid conclusions with minimum of effort andinvestment. The optimal product parameters may constitute International and NationalStandards while optimal process parameters may form by and large only Company or PlantStandards.

As stated above, there are a large variety of other Statistical tools to cater to variety ofother situations. These include decision to choose the product and location of the plant, choiceof equipment; choice of suppliers, their rating and development; production scheduling,inventory and store maintenance, scheduling despatches, assessment of customer satisfactionand what not.

The above unambiguously makes it crystal clear that any worthwhile Total Integratedand Synchronized Quality Management Programme for Continuous Improvement isincomplete without adequate dose of statistics.

Adherence to systems for conformance, to optimal developed standard, assures quality andyield delivered at right time and place at economically viable and competitive prices.



Lastly, it needs to be emphasized that there is one and only one unique optimal or rightquality for desired use to cater to specific market segment. A car or a motor bike run atunique optimal speed consumes least fuel, is more safe and comfortable on the road;minimises maintenance, service costs & pollution and simultaneously maximizes life ofvehicle. This way it benefits both direct and indirect stake holders.

Quality level designed to be better than the optimal might cost more and render ituncompetitive. While quality level worse than the optimal will not do. In either case, it is notpossible to sustain the market and the fall begins sooner than later leading to dissolution, theobvious logical end. Thus quality conforming to the rightly designed products and servicesbenefits all stake holders, direct and indirect, the society at large. It requires thoroughplanning, execution, review and updating. This cycle needs to be sustained eternally.

Secret of Japan to become global leader in quality lay in harnessing its mass medianetwork and institutional infrastructure to educate its entire human resource in the concepts,methodology and practice of simple quick and cost effective techniques. This enabled them tofully exploit the resources available to generate other necessary resources to create superiorproducts that endeared the people world over. Japan took two decades to achieve this status.India has the potential to lead the world in much shorter time now. It only needs to startworking earnestly.



2CAUSE AND EFFECT DIAGRAM

2.1 WHAT IS IT?

Let us first list, all parameters or characteristics of a product, component, service or anactivity of interest, which influence the satisfaction of the customer in particular and thesociety at large. Each one of these output parameters depends upon parameters of inputmaterials, processes, equipments and allied resources including jigs and fixtures,environment, skill of workers, measuring instruments and test facilities etc. Listing of allthese inputs called causes and its linkages with the outcome of concern such as yield, cost,non-conformity or excessive deviation called effect, when presented diagramatically is calledcause and effect diagram. Each effect has a cause and conversely each cause an effect, nothinghappens by itself. The law of karma is also the law of cause and effect. It is rooted in theuniversal law of as we sow, so shall we reap.

2.2 HOW TO DRAW IT?

Call a meeting of associated personnel representing Product Design, Process Design,Production, Inspection, Packing, Marketing and associated indirectly with the problem onhand or the effect of concern. From among this group a leader is chosen consensusly. Heconducts the brain storming session by introducing the problem and inviting suggestion forthe likely reasons or sources thereof. He presents these in the form of a diagram on thedisplay board and keeps on updating it. The final outcome of several sittings and attemptsmay look like the one's shown in Figures F2.1 and F2.2. While so doing, care needs to betaken to avoid any action that might discourage the participating members to express theirviews frankly.

Conducting brain storming session is an art. The minds of the participants needs to beopened up. The leader has first to open up himself and endear others. A leader once had toask the participants to forget about the problem on hand temporarily and instead enumeratethe various ways in which a glass tumbler could be used. The participants gradually openedup and contributed to the list of thirty six uses, against a normal enumerated number of 5 to10 ways. The problem was taken up again. The participants were found to contribute moreeffectively.

CAUSE AND EFFECT DIAGRAM 11


FIGURE F2.1 Cause and Effect diagram of nonconforming thickness of coating.

FIGURE F2.2 Cause and Effect diagram of slag inclusion in a cast iron product.

The attempt is to collectively fix the problem and not the blame. This is possible bythinking for what went wrong rather than who went wrong. Who went wrong or in whosepresence the error crept in, is vital, since he alone is in a position to tell how all it happenedand possibly how it could have been avoided or how it can be prevented in future.

Therefore a congenial atmosphere needs to be created where the team members areencouraged to share their experience in completing the diagram adequately. It needs to beexhaustive enough to include the culprit cause. Subsequently, it is these causes which shall beinvestigated for their contribution to the problem. If the culprit is not in the suspected or

Interpassroot

Fit up

Type

Type

Size

Skill

Technique

Manipulation

WelderConsumableProcess

Root

Design

Cleaning

SpeedArc lengthPolarity

Current

Slaginclusion

Weld joint Cleanliness

Coating agent

Foreign matter

Coating machine

Coating speed

Solid content

ViscosityRoll RPM

Roll bending

Nonconformingthickness

InstrumentThickness

Surfacetreatment

MethodMoisture

Substrate material Measurement



conjectured list of possible causes or the developed cause and effect diagram the solution shallbe evasive. The subsequent attempt to find the solution to the problem will lead to theconclusion that the root of the problem is not among the causes listed hitherto, but beyondthese. The whole exercise shall have to be repeated. Therefore adequate care needs to betaken to ensure that nothing vital is missed. Erring on the wrong side hurts less in the longrun. The cost of studying one extra harmless factor is negligible compared to the cost ofrepeating the study to assess the impact of the additional conjectured factor in conjunctionwith all the previous ones suspected to be interacting.

A few attempts shall be necessary to arrive at a fairly good and useful picture. Thediagram needs to be up dated as more and more experience is gained.

2.3 HOW SHOULD IT LOOK?

The cause and effect diagram should not look over congested or under congested. The formerpin points that problem is complex or large and needs to be split. See diagram F2.3 fornonconforming thermos flask that failed to retain the desired temperature for the stipulatedduration during final testing.

FIGURE F2.3 Cause and effect diagram of a thermos flask failingto pass the temperature retention test.

Each one of these sources can be split into specific cause and effect diagram. The one forsilver coating is shown in Figure F2.4

The latter is indicative of not enough thought having been given to the problem needinganother brain storming session probably by augmenting the team too, save exceptions. As an

Conformance Instrument

NormalNormal

PoorAbnormal

Method

Special S2

Special S1

Testingprocess

VacuumSpace

betweenwalls

Mis match

Uniform

UnevenLoose

Knockedout

Misplaced

Missing

Broken

Lidclosing

Silvercoating

Levelfilled

Size

Full

ThinSmall

Medium

Cylindricals

Verylow

Partialmarginal

Large

Poortemperature

retention



example see Figure F2.5, showing cause and effect diagram for preparation of tasty tea. It isnot adequate. The more complete version is shown in Figure F2.6.

FIGURE F2.5 Cause and effect diagram for preparation of tasty tea.

FIGURE F2.4 Cause and effect diagram of deficient silver coating, causing poor temperatureretention in a thermos flask, both with respect to its quality and consumption.

Shell weightspace betweeninner and outer

walls

Silvercoating

Excess

Short

Uniformity

Solution strength

Quantity deposited

AngleDuration

SpeedHighLow

Vacuumisingtemperature

Rollingprocess

Deficientsilver

coating

ProcessMaterial

Additives for mix

Sequence of Ingredients at variousstages and rateof increase

Temperature of

Duration

Water

Filtered

Form Sweetner

Ingredients

QuantityCeramic

MelmoBrassService

Metal

SteelCrockery

Tonguesensitivity

Thickness

SizeShapeHealth

UtensilsWeather Coopper

Earthen

SteelGlass

Brass

Size

Equipment Tea maker

Education

Oven & FuelTraining

MoistureTree source

Wood

DungExperience

TeatasteOrigin

Tea

Milk

Mood

Tea taster

Mineral

Chicory TulsiFlayour

Kitchen environment& hygiene

Adrak Lemon



FIGURE F2.6 Improved version of the cause and effect diagramfor preparation of tasty tea.

The diagram enhances grasp of the inter relationships among various causes and theirtotal impact on the effect under study. Besides depicting logical linkages, care is taken aboutits aesthetic look or appearance.

The main arrow should normally be bold and proceed from left to right. The effect shouldbe bold and prominent. It may be highlighted in a circle, square or an ellipse. The mainsources like man, material, machine (equipment), model (design), method (process) andenvironment are equitably divided between upper and lower halves. The arrows from thesetitles to the main central horizontal bold arrow should preferably be at 45 degrees. Thearrows for sub-causes under these heads can be horizontal. The sub-sub-causes may be linkedto these once again with arrows at 45 degrees. Incidentally such a lay out facilitates inclusionof more conjectured sub-sub-sub causes.

This diagram was introduced and propagated by Dr. K Ishikawa of Japan. This diagramis therefore also known as ISHIKAWA DIAGRAM. Again, if the written words are removedfrom the diagram the structure of the arrows resembles that of bones in a fish. Hence thenomenclature of FISH BONE DIAGRAM. Call it by any name it provides first successfulrational step in search of the solution to the problem. One needs to get into the habit ofstarting the solution to the problem with drawing its cause and effect diagram.

Adrak LemonMaterialProcess

Temperature of

ingredients at variousstages and rate

of increaseFlavour

Kitchen environment& hygiene

DurationReady

Mix

Separate

ServiceBrass

Metal

Steel

Crockery

Ceramic

Silver

MelmoForm

Ingredients

Quantity

Thickness

Sacchrin

SugarPowderCubes

SweetnerLump

Gur

CTC

Lipton

Brooke bond

Processor FormBagDustGranulesLeaves

Tea

Origin

Teataste

Experience

DungBuffaloCow

WoodCoal

SizeTree sourceMoisture

Oven & fuelTraining

Education

Tea markerEquipment

Size GasK. oilGlass

Brass

Steel

Earthen

Weather

Mood

Health Shape

Tonguesensitivity

Tea taster

Copper

Utensils HardSoft

Nilgiris

DarjeelingAssam

Kangra

MilkCowBuffalo

Goat TonedDaily

Form Source

FreshPowder

BoiledCondensed

Powder

Mineral

Sequence of

additives for mix

Chicory TulsiFilteredWater

Crystal



2.4 WHAT ARE THE POTENTIAL AND SCOPE OF APPLICATION OF THISDIAGRAM?

Its potential and scope is tremendous. It is universal. It develops a habit of deep thinking ofeach process or input that goes in to the making of the product or the problem and itscontribution to the same. The experience even leads to the knowledge of what factor hurtsand by how much. This knowledge contributes to further efforts to optimize the process andthe outcomes. A culture of drawing cause and effect diagram for every problem encounteredin an organization is an index of its growth potential.

2.5 SOME MORE EXAMPLES OF CAUSE AND EFFECT DIAGRAM

Figures F2.7 and F2.8 depict the factors influencing the outcomes of performance of a Teamin a game of Sport and National Status on Quality in the World Market, respectively.

FIGURE F2.7 Cause and effect diagram of performance of a team in a game of sport.

Capability to respondto tricks of oppoinent

Stamina

Will power

Attitude

Determination

Appropriateness

Exercise(s)

Duration

Ease

SpeedSkill Strategy

Never give up

Exploit weakness of opponent

Doing ones bestSpecialtricks

Concentration

General

Practice

Special

Practice

Rest

Nutrition

Strength

Mobilityof limbs

Extent

Regularity Performancein sport



Meetings

Review

Govt.polcy

Monopoly

Q. MarkCompetition

Standardisation

Q-awards

CurriculumLevel

NumberInstitutions

Mass

Nutrition

Frequency

QualityLong

termEducation

Audits & trainingprogrammes

Activites

Shortterm

Education Training

Incentives

Expert

visits

Managementpolicy

Law

Nationalqualitystatus

Professionalsocieties

Ammenties &facilities

Contacts

Commitment

Commitment

Responsibility

Organisation

Standards

system

Investigation

Company wide Q.CHuman resource

TeachersOrganisation

Educationsystem

Budget

Q.C.

circles

Seminars &conferences

Labourunions

Attitude

Sports

Health

Manuals

Consumer societies

Membership

Mass

movementPublications

Number of

branches Nature ofinformation

FIGURE F2.8 Cause and effect diagram for national status on quality in the world market.

CHECK SHEET 17


3CHECK SHEET

3.1 NEED FOR CHECK SHEET

A good quality control programme aims primarily at prevention of non-conformities. Itrequires generation of data on output parameter(s) and corresponding input parameter(s). Thedata thus generated are used to assess, whether what is being done is as required or planned.If however inspite of conformance to stipulated input parameter(s), the results are un-satisfactory, the data are analysed to determine new optimal level(s). Trial runs are made toconfirm the projections. These are then implemented if found satisfactory. Otherwise thestudies are repeated.

An appropriate standard format is necessary to record the information necessary to fulfillthe contemplated needs for exercising measures for prevention and improvement. This formatgoes by the name of a check sheet.

3.2 WHAT IS A CHECK SHEET?

The format for recording all the necessary information to execute preventive plans and toenable development of strategies for improvement, commonly known as Data Sheet, LogBook, Inspection Record, Schedule of Enquiry has been termed check sheet in this context.Such a check sheet needs to be standardized to enhance its utility. It is better to arrive at itsdesign in consultation with the users�who are to record the information, summarise andinfer for immediate use or action and review for in depth analysis for feedback forimprovement of control and to aid break through studies.

The nomenclature of check sheet has possibly emerged from the standard check listprovided to the packer to verify and ensure that all the accessories meant for an equipmenthave been accounted for in the intended package for dispatch to the customer or the tickmark resorted to by the accountants to confirm that the item thus marked during the auditundertaken by them has been checked and found to be conforming to the stipulated rules andprocedures.

As experience is gained the check sheets need to be continuously (periodically) upgradedand modified for improvements to serve the intended purpose more efficiently.



3.3 CONTENTS OF A CHECK SHEET

The contents of check sheet should be adequate enough to fulfill the contemplated needs ofprevention and improvement. The following is illustrative list of contents.

● Identity particulars to enable traceability like organization, date, project, customer,supplier, specifications, batch or lot number, as applicable.

● Associated human resource and Equipments if any● Input(s), implying consumption of resources of all kinds with pertinent details of their

quantity and allied parameters like purity or strength or size as the case may be.● Process parameters● Environmental parameters of concern if any● Method of sampling, sample size, frequency of sampling● Inspection and test facilities used or availed● Inspection and test method, least count or visual standards adapted● Corresponding output(s) with details of parameters of interest● Indices or appropriate functions of observations madeAs already stated it is an illustrative list only and by no means exhaustive to cater to the

diverse needs of huge variety of situations. Whatever is not considered worthy of the effortmay be ignored and any other pertinent information missed needs to be included.

3.4 SALIENT FEATURES OF THE CHECK SHEET

It should contain all the vital information. It is a common belief that often unnecessary detailsare asked for. A standard question will address these doubts. Will it or will it not be useful ininvestigations, that may become necessary in the event of a nonconformity having beenobserved internally or reported by the recipient, to identify the cause or source of non-conformity and find a fool-proof solution to the problem ? If this is likely to aid, it needs to beincluded or else ignored.

It should be easy to document the required information and the necessary facilities to doso should be available. The amount of writing involved is minimised. The layout shouldfacilitate summarisation, allied calculations and inference for associated decision making andfollow up actions.

The cost of developing and maintaining a standard check sheet should be viewed from thelikely losses, the lack of it might entail.

3.5 SOME ILLUSTRATIVE CHECK SHEETS

3.5.1 Attribute Inspection

A heavy electrical equipment manufacturer had a system in place to inspect, the fabricatedjobs accomplished by the production to their satisfaction, by an inspector belonging to anindependent inspection department. An inspection report of a huge fabricated job read as:

The job accomplished is shabby. The locations of non-conformities have been marked �X�.The job may be re-submitted for inspection after necessary compliance.

The production department re-submitted the job after doing the needful. The inspectorrecorded the following observations in his next report:

CHECK SHEET 19


It is false. The job has not been rectified in its entirety. The job needs to be re-submittedafter completing necessary rectifications.

Has the inspector communicated his findings adequately to enable the rectifier to performhis role satisfactorily for a huge job of this magnitude, perhaps exceeding ten cubic meterswith a large number of sections and segments? Locating all unspecified number of X�s itselfconstitutes an uphill task. It would certainly have been helpful to give the following additionalinformation(s) in that order:

● the number of X�s to enable the taskmarked on the job and performer to do the jobtheir locations without omissions.

The locations may be descriptive like, front middle box inside top and so on, or indicatedby an �X� mark on the attached pictorial drawing.

Certainly, any additional information on:

● the nature of non-conformity● the degree of its criticality indicating its likely potential harm● the likely source of the fault, of course may be

as opined by the inspector and given as a● the possible remedy (not documented foot note

in the work instructions in vogue) as ora measure of suggestion remarks

shall go a long way in planning and execution of the preventive measures.The lay out of the check sheet plays an important role in making a summary to enable

prioritization and strategic planning for necessary follow up actions.A possible check sheet in this context is illustrated in Table T3.1. This is fairly elaborate

and caters to a large number of similar situations. It is a three dimensional matrix. It can be

Table T3.1 An illustrative check sheet for an attribute data.

TYPICAL CHECK SHEET-ATTRIBUTE

Organisation: ‘O’ Department: ‘D’ Component: ‘C’ Drawing No: ‘DN’Date: Shift: Operator:

Tally marks for nonconformities (defects)Inspector:

Defectcode

Criticalitycode

Location code

D1

C1

L1

III III

I I

—

L2 L3 --Etc. Total

C2C3

Etc.D2

Etc.Total

List of defect, critically and location codes should be provided.N.B. 1. In simple situations location and or critical code may be redundant.

2. This check sheet is very appropriate for casting, fabrication, forging,ceramics and similar jobs.

3. Instead the job itself or the, ‘drawings’ of the component and itsvarious sections or views may be used to indicate the location of thedefect where the ‘defect code’ and its incidence can be indicated.



curtailed to address the needs of simpler situations adequately. In its present form, all thatthe observer is required to do to record his observation, is to put a tally mark in theappropriate �cell�. Row totals, sub-totals of criticality categories, column totals, sub-totals forlocations and grand total provide sufficient summary information for a job. Such respectivetotals over jobs enable identification of dominant sources to plan and execute steps for theiravoidance and subsequent control to sustain the improved status.

3.5.2 Variable Inspection

An electronic company engaged in manufacture of audio equipments purchases lots of tinycomponents of large sizes (numbers). Its acceptance sampling plans chosen according to IS:2500 call for samples of size fifty. The acceptance criteria and procedure require breaking ofthe sample into ten sub-samples of size five each; calculating sample ranges and means asalso grand mean; calculating further �statistics� for comparison with corresponding acceptablelimits to ultimately decide on its worthiness for acceptance or otherwise for conformance todesired target and spread. Instead, check sheet shown in Table T3.2 simplifies data recording.It straight away builds the Histogram (See Chapter 7). Its interpretation with superimposedtolerance limits provides clarity about:

● approximate capability of the process● process status in respect of central tendency and pattern of spread● acceptability or otherwise of the lot, as also● remedial measures required on the process to make it acceptable.

Table T3.2 showing illustrative check sheet for variable inspection.

TYPICAL CHECK SHEET-VARIABLE

(Identity and traceability particulars)

Dimensionclass interval

Tally marks Total

30.025 7-30.025.075 0.125

.125 .175LSL

USL

.175 2.225

.225 4.275

.275 10.325

.325 41.375

.375 25.425

.425 23.475

.475 39.525

.525 6.575

.575 0.625

.625 8

165Total

.675

N.B This is very useful for1. Incoming inspection2. To confirm setting target when it is fast moving job but likely to last

for short duration.

CHECK SHEET 21


For example the data of Table T3.2 indicates that:● the process is not centered at the desired mean of 30.325 and instead is bimodal

meaning that the process is mix up of produce of two distinct averages viz 30.350 and30.500.

● there are sporadic incidences of a few avoidable over size and under size components● the process seems to be capable of producing components conforming to desired

tolerances.

3.5.3 Packing Process

Machinery manufacturers often receive complaints that the spares accompanying the machinewere incomplete. They even put in a claim for the replacement and the damages caused bythe delay in the completion of the project. The remedy lies in preparing a standard checksheet (list) for each equipment. The packer then ticks each item in the list as it is packed andthus avoids complaints and avoids additional costs necessitated by extra correspondence,packing and dispatch besides compensating for the damages claimed. A possible specimen ofsuitable check sheet for such situations is shown in Table T3.3.

Table T3.3 A typical check sheet for list of items to be packed and or dispatched to a customer such asa standard list of spares accompanying a machine tool as accessories.

3.5.4 Work Sampling Study

Work sampling study consists of observing the equipments or even human resource atrandom (un-predictive) intervals of time; on whether they are busy, if so on what activity andconsequences of interest; or if idle, the source or reason there of. Such studies are aimed ataugmenting the efficiency of the system. The proposed check sheet for such purposes isillustrated in Table T3.4.

TYPICAL CHECK SHEET(Identity and traceability particulars)

List of items and quantity to be packed and ordespatched to a customer (say as spares)

S. No.

1.2.3.4.5.6.

N.B. Similar lists may be made for performing and or inspecting certain jobsor activities such as

1. Assembly of cars, tractors, turbines etc.2. Maintenance of buildings, machinery etc.3. Count down for space program with due regards to their sequence.

Item

FastenerScrew driverGasketWeightRubber washerKeysEtc.

Size Quantity



Machine or round number

OperatorNumber

12

.

.

.

3

Etc

Entries in codesWorking status if working:No feed, idle runningWorkng at half capacityNon conforming materialNon conforming process

Cause of nonworking status:No materialNo orders/planElectrical breakdown-waitingMechanical breakdown-waitingMaintenance electrical in progressEtc

WNPWIWH

WNMWNP

WNPNMNP

NEWNMWNEP

(Time)1 2 3 4 Etc

CHECK SHEET FOR WORK SAMPLING STUDY

Company: ‘C’ Department: ‘D’ Period: ‘P’ Superviser: ‘S’

Table T3.4 Illustrative check sheet for conducting work sampling study.

3.5.5 Consumer Complaints

One of the most important driving force for under taking projects for improvement by specialdesignated teams are the dissatisfied customers. Their feedback is crucial. Apart from whatthe customer may have to say on the subject, he needs to be guided to provide additionalinformation, that is vital for resolving the problem or incorporating the improvementsvisualized by him, to enable focus on the specific aspects for necessary development. A sketch

Table T3.5 A specimen check sheet for recording Feed back from Customers.

Nature of failure code

A B C Etc.

TotalSuggestion.

TYPICAL CHECK SHEET(Identity and traceability particulars)

Town Usagetype

Month/yearof manufacture

Consumer complaints for equipment

N.B. 1. Town describes environment conditions

CHECK SHEET 23


of check sheet to cater to this need is illustrated in Table T3.5. It needs to be modified to meetthe demands of specific situations.

3.56 Setting Approval

The prerequisite for process control is to start right. The settings for the process parametersought to be perfect or as much perfect as possible. The record of settings made will be veryhandy and useful for effective control of the process through use of appropriate process controlchart. A specimen of a possible check sheet for approval of setting prior to commencement ofproduction is shown in Table T3.6.

Table T3.6 A specimen check sheet for setting approval.

3.6 HOW MANY CHECK SHEETS TO PLAN?

The Cause and Effect Diagram shall provide adequate guidance for stages needing the use ofcheck sheet(s), its type and the details of its contents. As mentioned earlier, the check sheetsneed to be revised and updated in the light of experience gained and more so to fulfill therequirements of the relevant updated cause and effect diagram.

2 3 4 5 6 7 8 9 10 11 121

CHECK SHEET FOR SETTING APPROVAL

Organisation : Machine No: Product:

Characteristic: Specification: USL.....LSL.....TOL.......

Setting approval: (USL + LSL)/2 ± TOL/4

SerialNo.

Date Time ObservationObservation

Total Mean Remarks InitialsDeviationfrom

target1 2 3 4



4PARETO ANALYSIS

4.1 WHAT IS PARETO ANALYSIS?

Alfred Pareto an Italian economist, discovered during his research studies, that it is a law ofnature that wealth (income) is concentrated among the vital few. About twenty percent of theof the top few (the vital few) possessed about eighty percent of the total national wealth whilethe remaining majority of eighty percent were contented with the rest of the twenty percentwealth. This versatile law of nature is called Pareto law. This is so true, that twenty percentof the over 500 countries the world over share among themselves about 80 percent of the totalglobal natural resources while the rest eighty percent have to manage with the remainingtwenty percent innovatively. Like wise, of the about 28 states and union territories in India,only 5 to 7 account for the major share of 80 percent among themselves, while the rest sharethe remaining twenty percent. Even in sports, take for example cricket, it is only a few 2 or 3out of 11 member team who make most of the runs or claim major share of wickets. Even theoffice efficiency depends more on the committed about 20 percent human resource, on whomthe boss depends for timely and correct delivery. This law, jolly well, extends to thedistribution of non-conformities or the loss emanating from these, among its many possiblecauses or resources. The art of analyzing the data to identify the vital few sources of lossesemanating from lack of quality, is called Pareto Analysis. For illustration see Figures F4.1.

4.2 WHAT IS ITS USE?

To maximize the result of efforts and allied inputs to resolve quality problems, it isworthwhile to concentrate on the vital few sources for their avoidance or avoiding their effect,instead of dissipating the limited energy on the trivial many. This is synonym with the policyof divide and rule, except that there is no need to perform the unethical role of dividing. Thenon-conformities by law of nature are already so rooted among its causes or sources. It is we,the human resource, who need to be united against the inanimate non-conformity to, not onlyeliminate it but also to prevent it from recurring or taking its birth again.

Lest, the message is misunderstood, it needs to be reemphasized that we must alwayswork in unison as one cohesive team, if only we wish to succeed in our mission, whatever itmay be. Divide and rule policy may be fine in the short run but certainly disastrous in thelong run. We are welcome to do so, if we are selfish enough to care for ourselves and neglectour next generations. We therefore prosper at the cost of those whom we love most and forwhose sake we believe or pretend to be working. Let us be honest to ourselves.

PARETO ANALYSIS 25


FIGURE F4.1 Pareto Analysis. Concentration of nonconformities by types of problem (cause).

4.3 HOW TO DO IT?

Once adequate and appropriate information has been generated in the check sheet(s), thenon-conformities in the end product or service confronting us are classified according to theircauses or sources, since these emanate from the operational levels of the parameters whichquantify the respective inputs that are to be acted upon selectively.

The causes are then ranked in diminishing (non-ascending) order of their frequency ofincidence or contribution to the loss, if the harm from the different causes is uniform. Thefindings are then pictorially presented in the form of a bar chart, with the �ranked� causes onthe X-axis and the �percent contribution� on the Y-axis. The cumulative figures, cumulating to100, are shown as a curve on the same diagram. Reference may be made to Figure F4.1, onceagain for its comprehensive look.

4.4 PRECAUTIONS

Care should be taken that the contribution from the category of �Others� or �Miscellaneous�causes is indeed negligible and preferably does not exceed 10 percent. If it is in excess, it maybe split into the major among these and the rest.

On the contrary, if the �number one�, the most dominant is too dominant and too complexto be resolved, it should be suitably split or subdivided to get a hang over the manageablepart of the problem. The small success will pave the way for enhanced confidence andcapability for success in the next iteration.

It may be worthwhile to mention, at this stage, that the cause that ranked number twohitherto becomes number one for the next iteration and so the phenomena or the struggle forstep by step continuous improvement continues.

Cumulative

Individuals

Per

cent

occu

rren

ce

100 100

0

Y

A B C D E F GX

Casuse code

Num

ber

oftim

esof

occu

rren

ceov

era

year

A. ContaminationC. Wrong scaleE. Wrong specificationG. Others (Miscellaneous)

B. Poor adhesionD. Wrong colorF. Flaw in (other) document



4.5 A CONTINUING PROCESS

When the vital few have been successfully contained on routine basis, namely theircontribution is under statistical control, it is time to repeat the Pareto Analysis study. Thecauses which were considered dormant so far, become vital now and need to be attended onpriority. The war on non-conformities continues till the goal of ppm (parts per million) or ppb(parts per billion) nay zero non-conformity is reached.

If at any stage of the journey, it is believed that zero defect has been reached, thecompetition brings in new concepts of quality and customer expectations; may be visualappeal, longer life, lower cost, enhanced reliability, additional functions that again show up asnon-conformities requiring attention. The quality journey makes a fresh start.

Arise, awake and stop not till the battle on poverty is won. The prosperity shall be ours ifwe raise the �average� earnings and reduce the disparity or the �spread�. The short cutconvenient approach to reduce disparities by artificial regulations defeats the very purpose ofthe noble mission nobly intended.

4.6 CONTRIBUTION TO LOSS IS PERTINENT

Some times it so happens that the loss contributed by a particular non-conformity is manytimes over than some other type of non-conformity. In such situations the Pareto Analysisshould be based on proportional contribution to the Quality Loss by the non-conformity ratherthan on the basis of relative incidence or frequency.

4.7 SOME EXAMPLES

4.7.1 Chemical Plant

Pareto Analysis of concentration of non-conformities.The summary of number of non-conforming batches of an adhesive for one month�s

produce is presented in Table T4.1.

Table T4.1. Showing Nature, Frequency and Rank of Non-conformity.Organisation: ORG Product: PRO Month: MO/YR

Non-Conformity

Serial Number Parameter Frequency Rank

Number Percent Roman Alphabet

0 1 2 3 4 5

1. Colour 31 50.0 I A2. Composition 6 9.7 III C3. Contamination 10 16.1 II B4. Sticker 5 8.1 V E5. Testing 6 9.7 IV D6. Viscosity 4 6.4 VI F

Total 62 100.0

PARETO ANALYSIS 27


It give details of Nature (Cause) of Non-conformity, Frequency (number and Percent) andRank number with Alphabetical Code.

The above Table is recast in Table T4.2, after listing the nature of non-conformity in theorder of their respective ranks with alphabetical designations.

Table T4.2. Showing Rank, Percent frequency and percent causes.Organisation: ORG Product: PRO Month: MO/YR

Non-conformity

Cumulative percent of Percent FrequencySerial Number Rank Parameter successive parameters Individual Cumulative

0 1 2 3 4 5

1. A Colour 16.7 50.0 50.02. B Contamination 33.3 16.1 66.13. C Composition 50.0 9.7 75.84. D Testing 66.7 9.7 85.55. E Sticker 83.3 8.1 93.66. F Viscosity 100.0 6.4 100.0

The plots of Percent Parameters or causes on X-axis (column 3) versus percentcontribution, individual on Y-axis (column 4) in the form of a bar and cumulative on Y-axis(column 5) in the form of curve are shown in the diagrammatic form in Figure F4.2 known asPareto Diagram. It shows that resolving the first cause will bring substantial gains. Afterresolving this, the Pareto analysis needs to be repeated to sustain attempts for continuousimprovement. Possibly, the cause at rank two may acquire the status of rank one. The processis thus kept alive.

FIGURE F4.2 Pareto Diagram of nonconforming batches of an adhesive.

00 16.7 33.3 100.0

10

20

30

40

50

60

70

80

90

100

Per

cent

con

trib

utio

n (lo

ss)

6.48.19.79.7

16.1

50

66.1

75.8

85.5

93.6

100

Percent parameters (causes)

X

Y



It is too obvious to state that, if the remedy for any non-conformity down the line isknown, fool-proof system to avoid its recurrence needs to be put in place immediately ratherthan waiting for its turn indicated by its rank. Such exceptions only prove the rule.

4.7.2 Heavy Electrical Plant

Pareto Analysis of non-conformities detected in Turbine Blades.The relevant data are shown in Table T4.3.

Table T4.3. Showing cause wise percent contribution tononconformities among Turbine Blades.

Percent contribution to nonconformitiesSerial Cause code and Percent causes

Number description cumulative Individual Cumulative

0 1 2 3 4

1. A Linear Dimension 10 32 322. B. Profile 20 25 573. C. Edges 30 19 764. D. Surface 40 12 885. E. Material 50 7 95

6-10. F to J. Others (Misc.) 100 5 100

The plot of column 2 on x-axis versus column 3 on y-axis as histogram and column 2 on x-axis versus column 4 on y-axis as curve presents the Pareto analysis based on the respectiveindividual and cumulative contribution from various causes. This is exhibited in Figure F4.3.

FIGURE F4.3 Pareto analysis of nonconformities amogn Turbine Blades.

A B C

Cause code

Per

cent

occ

urre

nce

D E0

100

50

Y

X

A. Linear dimension B. Profile

C. Edges D. SurfaceE. Material F. Others (consist of cause)

PARETO ANALYSIS 29


4.7.3 Foundry

Figure F4.4 illustrates Pareto Analysis from a cast iron foundry.

4.7.4 Engineering Industry

The details of non-conformities observed among small components in a heavy industry withrespect to its frequency of occurrence are shown in Table T4.4.

Table T4.4 Data on nonconformities observed amongsmall components in a Heavy Industry.

Serial Cause code Percent Rank Rank wise percent nonconformities

Number and description nonconformities Cause code Individual Cumulative

0 1 2 3 4 5 6

1. A. Curvilinear dimension 15 II B 75 752. B. Linear dimension 75 I A 15 903. C. Surface 6 III C 6 964. D. Others 4 IV D 4 100

The plot of column 6 along y-axis versus cumulative percent causes namely 25, 50, 75 and100 along x-axis as shown in Figure F4.5 depicts the corresponding Pareto Diagram.

Additional information was later provided that each nonlinear nonconformity costs tentimes as much to rectify as it costs to rectify a linear dimension. The cost weights for othernonconformities were also ascertained. This impacts the priorities for finding the solution tothe sources of nonconformities. The revised summary is given in Table T4.5.

X0 A B C D E F G H

Cause code

100

50

Y

Per

cent

occ

urre

nce

A. Blow holes B. ShrinkageC. Sand D. ShiftE. Mis-run F. SlagG. Dimension H. Others

FIGURE F4.4 Pareto Analysis of nonconformities in a foundry product.



0

10

20

30

40

50

60

70

80

90

100

Y

Per

cent

inci

denc

e (n

on c

onfo

rmiti

es)

75

15

0 25 50 75 100

64

X

Percent cause

90

96100

Table T4.5. Data on nonconformities observed among small components in a heavyIndustry by causes and their proportionate contribution to loss from rework.

Serial Cause code Percent Loss Loss (Relative) Rank Cumulative

Number and description nonconformities weight Amount % loss %

0 1 2 3 4 5 6 7

1. A. Curvilinear dimension 15 40 600 65.5 I 65.52. B. Linear dimension 75 4 300 32.8 II 98.33. C. Surface 6 2 12 1.3 III4. D. Others 4 1 4 0.4 IV 100.0

Total 916 100.0

Notes: Causes C and D together contribute less than 5 percent loss. It is therefore Desirableto club these.

The loss from cause A is too dominant. It may have many sources or subcauses. Therelevant details are desirable to focus on manageable area in first iteration.

The revised version of the Pareto Diagram appears in Figure F4.6. This gives a picturemore close to reality than the one already seen in Figure F4.5.

4.7.5 Service Sector�Telephone Exchange

The cause wise data on inoperative telephones are shown in T.4.6, both by the frequency of itsincidence and the duration of its remaining nonfunctional. In each category these are rankedin non-ascending order of their contributions.

FIGURE F4.5 Pareto diagram by incidence of nonconformitiesamong small components in a heavy industry.

]

Percent causes

PARETO ANALYSIS 31


A B C D0

50

100

X

Y

Cause code

Per

cent

loss

FIGURE F4.6 Pareto diagram by percent loss contribution bynonconformity types among small components in a heavy industry.

Table T4.6. Fault Code of Inoperative Telephone Instrumentwith frequency of its incidence and idle duration.

Telephone Exchange: TE Period: PE

Incidence�Number of complaints Idle duration-hours

Rank Fault code Number Cumulative Percent Fault code Hours Cumulative Percent

I A 1490 1490 30 C 7539 7539 32II B 1005 2495 50 D 5867 13406 57III C 954 3449 69 B 4452 17858 76IV D 883 4332 87 A 3261 21119 90V E 507 4839 97 E 1432 22551 96VI F 102 4941 99 F 458 23009 98VII G 51 4992 100 G 246 23255 99VIII H 9 5001 100 H 117 23372 100IX J 6 5007 100 J 93 23465 100

Fault Codes: A : NFF External B : Cable C : ExternalD : Subscriber�s Apparatus E : Subscriber�s Fittings F : NFF ExchangeG : MDF H : Hot J : Exchange

The respective Pareto diagrams are shown in Figures F4.7 and F4.8. The consolation isthat either way the first four faults remain the same. Thus for customer satisfaction, the firstfour causes accounting for a major hurt to the customers and loss to the TelephoneDepartment, need be addressed on priority.

It is pertinent to emphasise at this stage that the repeated use of the three tools, describedhitherto, in this sequence helps to diminish the size of the problem. We need to iterate till we



are in a position to crack the problem. To optimize the fruits of our efforts, we need to bite asmuch as we can chew. To begin well is a good strategy. Well begun is half done. Our firstiteration deserves preferential consideration.

C D B A E F G H JX

Y

10

20

30

40

50

60

70

80

90

100

0

Fault type

Cumulative

Per

cent

dur

atio

n in

oper

ativ

e

FIGURE F4.8 Pareto diagram of data of Table T4.6—Inoperative duration.

Pareto Analysis of Fault Occurrence in Telephones—Inoperative Duration Telephone Exchange : TE; Period : PE

FIGURE F4.7 Pareto diagram of data of Table T4.6–Incidence of occurrence.

Fault Codes: A : NFF External B : Cable C : ExternalD : Subscriber’s Apparatus E : Subscriber’s Fittings F : NFF ExchangeG : MDF H : Hot J : Exchange

A B C D E F G H J0

50

100

Y

X

Fault type

Cause code

Cumalative

Per

cent

freq

uenc

y

Pareto Analysis of Fault Occurrence in Telephones (Type wise)–Frequency Telephone Exchange : TE; Period : PE

STRATIFICATION 33


5STRATIFICATION

5.1 DEFINITION

Stratum has been defined as a layer or a set of successive layers of any deposited substance.Strata is the plural form of stratum.

Stratification is the process of classifying the data into groups such that the groups are asmuch homogeneous as possible �within� and heterogeneous �between�. Thus inter groups orstrata disparities or variation may be large but within or intra small.

5.2 UTILITY

This technique is used in socio-economic and allied surveys to obtain maximum efficiency inestimating appropriate indices. It aims at reducing sample size and simultaneously the errorof the estimate. In this manner, it minimises the use of resources besides reducing theduration for completion of the task on hand. All these together contribute to substantialsavings of the exchequer.

5.3 PRE-REQUISITES

In industrial and service sector applications, the pre-requisites for deriving maximum benefitfrom the application of this concept are:

● Identification of the quality problem through Pareto Analysis.● Drawing of a Cause and Effect Diagram through �Brain Storming Session� among

personnel of all disciplines associated with product or activity directly or indirectly andallied stake holder; that is, listing or enumerating all factors that influence the qualityparameter of product, service or activity under study; and

● Developing suitable check sheet for recording true data on quality parameters withtraceability to the corresponding operational level of each factor including all pertinentinputs.

If above stipulated conditions are satisfied, then from among all the seven simple quickand cost effective techniques, the Stratification costs least and delivers most. In fact, if thedata are not amenable to stratification by the levels of factors, conjectured to influence the



effect, the parameter of the output of concern, the data are useless. What data to collect, howmuch to collect, to what accuracy and the manner of recording needs to be planned optimally,for mission to succeed timely and economically.

5.4 PROCEDURE, SALIENT FEATURES AND POTENTIAL

Once the data have been collected, the stratification approach implies, classification of eachquality parameter by levels of one or more factors at a time depending on total amount ofdata and technological knowledge of their interdependence.

Quality is known to be influenced by design or model, process or method, equipment ormachine including allied jigs & fixtures, location & environments, testing equipment includingallied facilities and instrumentation, process parameters, human resource or worker,maintenance, sequence or time of production etc. The term �Human resource or Worker� hasbeen used in a very broad and general sense. He may take the form of operator, inspector,helper, packer, user or a team. Thus quality parameter of interest is classified according tothe levels of each of the above relevant conjectured factors of influence, one or more, if theyare interdependent, at a time.

5.5 SUCCESS STORIES

5.5.1 Stratification with Respect to Workers

(a) Vanaspati (Hydrogenated Oil) Production Unit

A vanaspati production unit was manufacturing its own 16.5 kg tins for packing, in itsauxiliary unit. About 16 percent of the tins were found to be leaking during one hundredpercent post production or pre-filling inspection, as a measure of control to avoid subsequentlosses. Manufacturing operation consisted of stamping top and bottom square pieces; cuttingsheet, shaping it to cover all the four sides and seaming the side to make a hollow cuboid;seaming the hollow cuboid with the top and the bottom, followed by soldering on the seamedside.

Examination of past one month�s data revealed steady performance of rework of about 16percent as stated earlier. It is therefore an apt problem of Break Through. A dialogue withthe inspectors further revealed that the non-conformity of leak was localized and mostlydetected on the side seam. This by itself, is an illustration of informal geographicalstratification. This operation is operator dominant. It was therefore decided to study thefailures operator wise.

Six workers were performing the soldering operation. The worker (soldering operator)wise classified summary disclosed that, the rework was almost �nil� for 5 of them and almostthe entire rework was emanating from the 6th one only. The examination of the 6th workerdisplayed the obvious. He had a deformed thumb. He was assigned another suitable job anda trained substitute worker remedied the situation. The savings in the form of resultingnegligible inadvertent re-work, elimination of re-handling and enhanced production are worththe small effort made to remedy the situation.

STRATIFICATION 35


(b) Tin container manufacturing unit

Another factory, producing tin containers of the same size as in (a) above, was supplier tothe producers of hydrogenated oils. It was facing a peculiar problem of �non-acceptance� of thelots by the vendee. A check inspection at the vendee�s site confirmed prevalence of about 10percent leaky containers. The chief of the manufacturing unit was amazed, since he hadimplicit faith in the competence and loyalty of his inspection team engaged in doing onehundred percent pre-shipment inspection check of the tins. He therefore, conjectured thepossibility of damage in transit causing the leaks. If the conjecture of the chief, that thedamage to the empty tins occurred during transit of short distance to the customer�s site weretrue, the consequences of damage and resulting leakage loss and accompanying mess duringtransit of filled tins by the customer to his customers could be disastrous. A sample check, oftins ready for dispatch, confirmed prevalence of ten percent leaking tins and established thatthe stage of inspection was indeed the weak link in the system. A sample check of the outputof each inspector was done. It revealed about 30 percent evading detection for three of theinspectors and practically �nil� for the remaining seven. It so happened that these threeworkers were union leaders. They enjoyed the confidence of all. They were blissfully ignorantof the ground reality. Incidentally, they were giving more output and enjoying productionbonus too. These workers became conscious and cautious, on knowing the problem and itsrepercussions.

Never the less, the output of the inspectors was stacked separately and samples werechecked for confirming the �zero leakage� status prior to dispatch. There was �nil� complaintfrom the customer for next six months, thus eliminating the problem. This saved the businessfor the good of all!

Control of manufacturing process, including improvement if any called for, to produce zeronon-conformity that will avoid the existing necessary evil of one hundred inspection plussample check would be better alternative.

(c) Pharmaceutical Company

In one of the bottle filling and packing units of a pharmaceutical company, four workerswere engaged in identical operation of labeling. Output of one of the workers was twice thatof the rest of the individuals. The movement pattern of the hands of these workers wasobserved for a short while. The pattern of movement of their hands differed, leading todifferential units of work or energy (ergon). The science dealing with such aspects is called�ergonomics�. The best worker was promoted to train all the workers, not only in this unit butalso in all other filling and packing units. He was empowered to reorganize these suitably toenhance the overall performance. Thus a new standard procedure was established and put inplace for this kind of work. This approach paid rich dividends.

(d) Electrical Transformer Manufacturing unit

In a plant manufacturing oil transformers, it took a team of 4 to 8 workers to assemble aunit in 2 to 8 hours depending on the model. All transformers thus assembled are checked forbeing leak proof. The rectification process, in the event of a leak being detected, took couple ofweeks. Thus the �nuisance value� or tangible and intangible losses together caused by atransformer failing to pass the leakage test, are too obvious to be spelt out.

The average rework was about 22.5 percent. There were several teams or groups ofworkers performing similar tasks. Team wise stratification of their performance for past sixmonths was accomplished. It was found that two of the teams delivered one hundred percent



conforming (non leaky) transformers. Based on this part of the information, true toexpectation, the performance of the worst group was as bad as forty five percent non-conforming (leaky) transformers.

Detailed discussions as a part of the investigations and attempts to find the reasons thereof, in search of the solution to the problem, it became known that the leading pair of teamshad achieved �zero non-conformity� by assembling components thoughtfully selected from theavailable stock. Workers of the other teams assembled the left over components which lackedfitment criteria. They did not spell out this fact, for reasons best known to them. One of thereasons could be that there was no incentive for good performance by individuals nor anydisincentive for poor performance.

The remedial measures rightly concentrated on development of supplier(s),standardization of packing, transportation, preservation and handling practices to ensureinput of components of right quality for assembly.

(e) Export Inspection Agency

Export Inspection Agency Inspects goods on sample basis and certifies it export worthy, toprevent complaints from abroad and to build reputation for Indian products. Inspite of this,there are complaints. The inspectors get benefit of doubt on the basis of sampling errors asalso on the factors presumably beyond their control after they have certified these to conformto stipulated standards. A two way table of �sufficient� data classified by �inspectors� and therecipient making a complaint revealed the true story, of course after providing for �reasonable�benefit of doubt arising from sampling errors. It identified the dispatching and receivingstations lacking integrity.

To sum up, these are glaring examples of study teams getting at the root cause ofproblems to eliminate these. It simultaneously avoids mutual infighting on hunches, to blameeach other for the survival of the individual. Alternatively, the existing culture of poorperformance survives and havoc continues. Team work, instead, eliminates the problem forevery one to survive.

Generalising, the meaningful stratification with respect to workers, helps in identifying thebest, moderate and poor workers. The observation of a sample of workers from each of thesegroups generates the knowledge of DO�s and DON�T�s for performing the task on handefficiently. The establishment of standard methods and training in their adoption yields manybreakthrough results in a large variety of situations in all kinds and sizes of industries andservice organizations.

5.5.2 Stratification with respect to machines

(a) Chemical�Rayon pulp production unit

A rayon grade pulp manufacturing unit had an elaborate system of input and outputparameters at each processing stage, namely, chipping, digesting, bleaching and papermaking. However, no purposeful analysis of data was done for any effective action forimproving quality or augmenting production. The inconsistent quality of chips in respect of itssize, in the chipping house, was adversely affecting the subsequent processes. It was safelyattributed to heterogeneity of the quality of incoming bamboos from the forest, classified asdry, green, rotten or yellow.

A simple tailor made special study, called work sampling, was organized. Each chippingmachine was observed at random intervals of time. The quality of bamboos (dry, green, rottenor yellow) being fed into the chute of the machine and the quality of chips (size) being

STRATIFICATION 37


produced were recorded. Supplementary information gathered consisted of the number ofbamboos in the chute of the machine, the condition of the blades and its fitness. Theexamination of data thus obtained revealed that the quality of chips in respect of size, wasboth conforming and non-conforming irrespective of the quality of bamboos fed. Furtherinvestigation revealed that the cut quality of chips depended more on the sharpness of thechipping blades loaded on the machines and their adjustment. In depth scrutiny disclosed thatsharpness angle of the blades and their adjustment on the machines influenced the intervalfor re-sharpening as also its life. These have direct bearing on the economy of chippingoperation. It also highlighted the need to monitor the feed of the bamboos. It was seen to bevery very erratic. Any shortage or excess in the number of bamboos being fed directlyinfluence the productivity and life of the blades.

This is an illustration of rewards of stratification with respect to machines. It belied themyth that the fluctuations in the quality of chips with respect to their size were caused by thechanges in the status of bamboos received for chipping. The myth prompted, no action. As aresult, heterogeneous chips were fed to the digesting house. It created problems for thesubsequent operations. A machine wise periodic check, of input and output, was introduced tomake adjustment or replacement of blades, as a part and parcel of process control procedure.

(b) Electrical Engineering�Filling of cylindrical shells and sealing caps in assembly ofcapacitors

Two identical machines were in use for assembly of capacitors. Their day to dayperformance was satisfactory. The confidence gave place to complacency and hence processcontrol procedures if any were being ignored. One day the customer�s order could not becomplied with. The production was short and the quality was poor. As usual a fire fightinginvestigation started.

It found that, one of the two machines was running smoothly. Its output was normal andthe finish of the components was excellent. The output of the other machine was less. Thefinish of its component was poor. The component would often get jammed. The operator wouldbe required to extricate the jammed component before restarting the machine. One could seea heap of crumpled shells (scrap). He was under pressure to produce to meet the demand forthe day. The machine was in need of repair.

The pros and cons of doing the needful for the machine before commencing productionversus running it for the day are too obvious to be taken lightly. The company paid a heavyprice for giving a go by to the maintenance. The price included the cost of segregation, partialcompliance of the order and more damage to the machine adding to the cost of inevitablemaintenance and associated intangible loss of customers goodwill.

5.5.3 Stratification with Respect to Time

(a) Textile Mill

Generally, in a textile mill, the productivity during day shift is better than in the nightshift. However, in one of the mills, the converse was found to be true. The investigations leadto the identification of healthy practice of planning production. All small and special jobs,likely to need attention from superior authorities, were deliberately executed during the dayshift. This enabled timely aid, from the right quarters, on all matters needing attention.Hurdles, if any, in the path of production were taken care of. All the remaining hassle freejobs were planned for the night shift. This was the secret of enhanced performance during thenight shift. So much so, that the total conforming output for all the three shifts was way



ahead of other competitors. This even demystified the theory that when the cat is away miceplay, implying that in the night shift the work is lax.

(b) Environmentally sensitive industries

Classification or stratification with respect to shift also brings out the effect ofenvironmental temperature on the product, provided the job is not labour intensive.Alternatively, the differences are the combined effect of �shift environment� and the skill &stamina of the worker performing the task. The contribution of each can be assessed byclassifying the data simultaneously with respect to shift and worker or group of workers, asthe case may be. In some, situations even the day of the week has its own influence. In suchcases three way classification is called for. Often fresh data are not necessary. The past datasuffices. Workers are routinely changed in shifts, often weekly. Past three weeks data areadequate. If necessary, it may be collected a fresh. Three way classification of data by threeshifts, three weeks and three workers or teams, provide very useful assessment of the effectsof respective factors being sought to resolve the problem. In addition the differences amongthe averages for the six days of the week provide the measure of the systematic differencesamong days, if any and sow the seeds for developing strategy for improvement.

(c) Engineering Industries

Stratification with respect to time clearly brings out the effect of wear and tear of tools toguide, the formulation and adoption of standards for intervals and magnitude of adjustmentscalled for. This enables optimal use of tools and simultaneously assures production of desiredquality right first time.

(d) Chemical Industries

Similarly in chemical industries, stratification aids in assessing the rate of fall in strengthof chemicals like pickling and plating baths to enable formulation of optimal standards fortimely additions and replenishment inclusive of their quantities for adoption. This alsoassures quality right first time. It makes optimal use of chemical and reduces rework toenhance efficiency.

Stratification with respect to time from the commencement of the job gives signals for anyappreciable deterioration, particularly for time dependent activities or jobs, before it is toolate.Timely action avoids subsequent hazards. A stitch in time saves nine.

5.5.4 Geographical Stratification

(a) Pharmaceuticals Marketing

A pharmaceutical company was satisfied with its growing sales and accompanying profits.The sales were stratified region wise. It revealed that growth rate was different for differentregions. Not only that, in some cases, the sales had declined. Sales of each region was thenclassified by products to identify the dominant products contributing to this adverse situation.Together these constitute, two way classification. This helps in going into causes for the fallmore specifically. May be, new substitute products have been introduced by the competitors orweather changes have influenced and diversified the demands or something else hashappened. In either case, the discussions, with these findings as the basis, shall guide theapproaches for alternative remedies to make up for the shortfalls in the sales of individualproducts in specific regions and enhance the total market share. In fact, the discussions leadto policy of aggressive sales in virgin areas.

STRATIFICATION 39


(b) Shoe Manufacture

A large shoe manufacture company took a policy decision to prolong the life of the shoesmanufactured and marketed by them. They introduced a scheme to offer concession tocustomers exchanging their old shoes for new. This ploy helped the company to collect a largenumber of old worn out shoes of various brands of their competitors besides their own. Afairly large representative sample from among these was examined for locations (geography ofthe shoes) that had worn out, rendering it unusable. This location wise stratification was usedto identify the location prone to fastest wear ; the next location, the 2nd fastest and the likelygap between the two, and so on. This information was useful in developing alternativestrategies. The final decision, on prolonging the life of the shoe depended on the estimatedadditional cost versus extra life possible and the resulting economy depending on the worththat the customer is likely to attach. What a strategy to win over the market?!

The concept is equally effectively applicable to component dominant assemblies likewatches, automobiles and compressors; as also foundry and fabrication. Innovative use of�geographical� concept is necessary to exploit its full potential.

(c) Instrumentation

The factories manufacturing auto meters or water meters and the like do one hundredpercent inspection to identify the non-conforming meters. These are then rectified andrechecked before clearing for marketing. The stratification with respect to componentsrequiring adjustment, repair or replacement indicates the specific components that requireimprovements. Each one of these is then taken up as a problem for solution by the respectiveteams. Many factories have made use of these strategies to more than double theirconforming production with the same set up and resources in a short span of time withopportunities for several more such iterations.

(d) Electrical fan manufacture

A ceiling fan manufacturing unit was doing one hundred percent inspection for itsconformance to noise, speed and wattage. The roof had provision for nine hanging positionsfor inspection. During inspection care was taken to use the same set of Blades, Condenser andTachometer. Voltage too was stabilized. Classification of non conforming fans with respect tothese positions revealed position biases in respect of speed. The nine positions in a squareroom were, as shown in the lay out in Figure F5.1.

Same fan at Position P5 will give maximum speed; at Positions P2, P4, P6 and P8medium speed; and at the remaining four positions at the corners namely P1, P3, P7 and P9the lowest speed. Thus the same fan had highest chance of acceptance at P5, lower at evenpositions and still lower at the remaining odd positions. This problem was resolved byassessing the biases and allowing corrections for the same. The results at position P5 weretaken as the acceptable standard.

Differences among analysts and laboratories are not uncommon. These arise from non-standard practices and dispersions among testing facilities such as equipments or instrumentsincluding materials and chemicals used for testing the item. Such practices need to bestandardized to satisfactory levels so that the total errors are harmless.



P7 P8 P9

P4 P5 P6

P1 P2 P3

Figure F5.1 Showing layout of the fans for inspection on the ceiling.

(e) Service Sector�Food Grain Ware House

A national food grain ware housing body as usual engaged in procurement of food grainsin areas that have surplus or abundant production and moving the same to the needy areasthat are deficit. There were some losses in transit. The problem was to estimate the same andalso identify the areas or routes that are more prone to such losses.

A sample study was planned to observe marked bags at selected dispatching and receivingstations (geographical locations). The weights dispatched and received were recorded. Inaddition a questionnaire was sent to these locations, seeking information on the weightdispatched by the sender and the weight received by the recipient. This information wassummarized in a matrix, a two way table, illustrated in Table T5.1. This, as mentionedearlier, is known as two way classification.

Percent weight received. Excess (+), Short (�)

Recipient Despatch Station Overall

Station A B C D E

V

W

X

Y

Z

Overall

Table T5.1 Showing excess or short weight reported.

STRATIFICATION 41


By looking at the last column and the last row, one can comment on the integrity of thesender and the receiver in recording of the weights dispatched and received by themrespectively, in addition to comparing the proneness of the routes with regard to losses intransit. The findings of the physical survey and the response to the questionnaire were inagreement. This added confidence in the conclusions drawn.

(f) Service Sector�Laundry attached to Hospitals and Hotels

The laundries practice standard procedures for washing linen received from differentlocations (wards or rooms, kitchens or banquets as the case may be). Depending on thecapacity of the machines and the models, standard regime for quantity of detergent,temperature, speed and duration are followed. There are occasions when the washed lot is notconsidered sufficiently clean and either it is given extra run or re-washed. Thus the capacityis under utilized. Instead, the procedure should not mix up the linen from different locations.In other words, the linen is stratified by locations. Each stratum is then treated on its meritor needs. The standards in terms of quantity of detergent per kg of linen, temperature,duration and the like should be developed for each source depending upon the nature ofspoilage. This did help the laundries attached to hospitals and hotels to improve theirperformance. It resulted in better customer satisfaction, increased life of the linen andreduced cost through improved utilisation of the installed capacity and allied resources.

(g) Fabrication, forging and foundry

In fabrication, forging and foundry operations, the two way classification by the nature ofnon-conformity and location is of immense value to the chemist and allied technocrats to pinpoint the likely few causes or sources of the non-conformity. These are then examined toidentify the culprit for corrective measures. The leading industries are successfully exploitingthis approach. Can others do without this?

(h) Pharmaceutical�Tablet Making

Homogeneity of the ingredients of the tablet and its weight are important parameters.The former, even if achieved during its preparation, is likely to get distorted, particularly

in unani medicines in powder form, during tablet making. The vibrations of the hopper, causelighter ingredients to move up and the heavier to move down. Thus the ratio of theingredients gets distorted and is not same for different layers or strata of the contents in thehopper. The remedy lies in avoiding over filling of the hopper, a practice often resorted by theoperators as a matter of convenience. Depending on the nature of the mix, a standard heightto which the hopper can be loaded needs to be developed and stipulated for conformance. Ifthis is not done, even the weight of the tablets will deviate from the intended norm, and sowill be the ratio of the ingredients. The tablets will not provide the standard dosage or theeffect to the patient.

In case of multiple punch tablet making machine, the periodic samples of a fewconsecutive tablets are taken to verify that the weights and their dispersions conform to theprescribed norms. This procedure is good enough, if the assumption that all punches aredelivering the same average weights, holds good. Unfortunately, often this is not so. In suchcases, the data thus obtained, are likely to show more dispersion between tablets of the samesample. This in turn over estimates the process variation or its capability. It may show up theprocess as incapable, unless relatively the tolerances are wide enough. Alternatively, theprocess may appear to be in control, when in fact it is not. If this is the case, it is indicated,



either by the runs of points (see chapter 8) on one or the other side of the standard line ORby the concentration of points close to the standard or the average line. The remedy in suchsituation consists of special check of each punch for any bias for over weight or under weightbeyond acceptable norms. This should be periodically re-affirmed. The periodicity depends onthe degree of the duration for which the calibration remains stable. Also, the samplingmethod needs alteration. A sample of a few consecutive tablets need to be taken from thesame punch. This is punch wise stratification. This method of sampling is possible, only afterthe revolving table of the punches has completed the round or the cycle. The number ofrounds shall equal the number of tablets in the sample. The data thus plotted on the chart ismore likely to show up punch biases, assuming that the process has remained fairly stablebetween successive samples. Yet, another option could be to treat a set of all tablets from oneround as one �unit�. A sample of at least two units is called for. The sample statistic should beadjusted or corrected for the number of punches on the table.

5.5.5 Stratification with Respect to Supplier

(a) Vanaspati (Hydrogenated Oil) production unit

In spite of conforming to stipulated regime, the end product did not conform to the laiddown standards. Production department blamed it on the laboratory responsible for testingand acceptance of the oil purchased, and the latter in turn blamed the former. The testrecords were being maintained supplier wise. The non-conforming batch of vanaspsti could notbe traced to the supply because the oil received from different sources would get mixed upwhen transferring to huge storage tanks. All test records showed that only conforming goodoil had been procured. Yet, the problem was there for all to see.

The higher authorities advised them to resolve the problem together instead of blamingeach other. It was agreed to trace it to the storage tank with identity of the suppliers of itscontents. A team was formed to over see the efforts and provide necessary support. The pastrecord was criticality examined tanker wise, with details of the suppliers, whose suppliesmade up its contents. It took some time to lay finger on the supplier, who was common to thecontents of the storage tanks, the oil from which resulted in non-conforming batches.

It was decided to keep a special eye on the future supplies of the suspicious supplier. Asthe next supply arrived, the seal of the tanker was checked. It was found to be intact. Thesample of the oil, taken from the exit pipe as usual, was also found to be satisfactory on test.However, the doubt lingered on. The truck (tanker) operator was asked to break the seal andopen the lock for inspection of the tank. He refused on one pretext or the other, such as, hehad instructions that seal must not be disturbed in any manner whatsoever, he did notpossess the key. He was over ruled. He was promised compensation for the loss caused bybreaking the seal and the lock. What a surprise ! The tank was partitioned into two halves,interconnected at the bottom. The sample from the other half was tested. It was found to behighly adulterated. This in turn made the entire oil of the receiving tank, unfit for productionprocess envisaged on the basis of the available test reports. Any production from this tankerwould yield unsatisfactory result. Thus the remedy was known. The unscrupulous supplierwas black listed, for this deliberate mischief.

The worth of maintaining records, that can enable tracing of the occurrence of non-conformity to its source, can be recognised, at least now. I S O : 9000 stipulates as one of itsmajor requirements. The details need to be good enough for the purpose, neither too muchnor too little. The former adds to the cost while the latter defeats the purpose.

STRATIFICATION 43


(b) Electrical fan manufacturing unit

The entire range of components needed for assembling a complete fan are painted. Thequality of paint is judged by the glossiness, surface hardness and whiteness it imparts as alsothe surface area covered per litre of the paint. The factory was maintaining satisfactoryrecords that facilitated traceability of the finished quality to the brand of the paint used andhence the supplier. The organization made use of this data to choose and patronize the rightsupplier(s). The price and the adherence to delivery schedules by the supplier were givenadequate due consideration.

(c) Rayon grade pulp manufacturing unit

The V belts are used on motors to drive the mechanism that cuts the bamboos into chips.The record of the brands of the belts used and their respective lives were maintained. Theexamination of the data revealed that one of the three brands which cost ten percent morehad a longer life by twenty percent. The economic gain of this brand is evident. It needs to bepointed out that even if the cost of this superior brand were twenty percent more or stillmarginally higher, it was still worth it because of more gains accruable from ten percent lesstime and effort needed for the replacement of the belts. In addition corresponding loss inproduction is avoided. With this kind of justification, none can raise a finger on the choice ofthe supplier who is apparently costlier. This experience, hits the nail right on the head. Itcautions against the blind policy of accepting materials on the basis of lowest quotation alone.

5.5.6 Stratification with Respect to Customer

(a) Electrical fan manufacturing unit

The term customer is used in a broader sense to include dealers and sub-dealers. A fanmanufacturer, reputed for its quality was suddenly flooded with returns and complaints. Thereturned fans were thoroughly inspected and dismantled if necessary, to verify the veracity ofthe complaints and the faults reported. Some of the fans were in excellent condition and someothers appeared to have been intentionally tampered with. These were not normal damagesexperienced in transit or handling. The returns and complaints were classified by regions andwithin a region dealer wise. The reason was traced to a dealer who had shown high sales toclaim more incentive or bonus. Not only that, he had even let out the fans on hire, made easymoney during the season and later returned the fans, as of poor quality, which on the top ofit had been procured on credit. The organizations try to identify and choose right suppliers. Itseems now that, conversely, the suppliers practice the same principle to deal with the �right�organization. The policy to fool the opposite party can prove fatal in the long run.

5.5.7 Stratification by Process Stage

(a) Pharmaceutical company

A pharmaceutical company engaged in manufacture of a baby tonic suddenly encounteredthe problem of the presence of glass speks in the sealed and labeled bottles during finalinspection. This non-conformity was too serious to be ignored. All bottles were thoroughlyinspected and only bottles free from speks were marketed. The team designated to resolve theproblem met and discussed the issue. The majority view was that the weak neck of the bottlewas leaving speks during the stage of sealing operation for the cap. Other conjectures pointedto the possible originating sources as hardness of cap and the process of washing besides host



of others. The study consisted of following a sample (batch) of bottles at all stages of theentire process - starting from their receipt in the stores through opening and cleaning,washing, filling, sealing to labeling and packing. The bottles were inspected at each stage forpresence or otherwise of the glass speks. The inspection procedure used magnifying glassesand the area was adequately lit. The washing stage was identified as the factor contributingto the problem. The washing process was observed minutely. The bottles rolled over therollers and the brushes inside the bottles, all immersed in the hot water container. Themetallic end of the brushes was hurting the necks of the bottle to produce glass speks. Thesein turn were getting into the bottle due to centripetal force. The design of the brushes wasmodified. The metallic part of the brush was covered with rubber sleeve. This eliminated theproblem.

5.5.8 Conclusion

The variety of above success stories substantiates the claim of the tools being simple, quickand cost effective. The application of the concept of stratification, among these, is simplest,quickest and costs least. This tool is versatile and universal in application.

Its application on the past data should be first step in the direction of reaching thesolution to the problem. It gives a head start. It helps in identifying the likely dominantfactors and their levels with favourable and adverse effects. One can choose the rightcombination, accept this as the standard and adapt it to reap the expected gains. The least itdoes is, it carries the study a step forward in search of the solution to the problem. Thestandards need to be revised, as more and more experience is gained. In addition to itstangible returns, the intangible rewards include the boosted moral of the human resource.

Lastly, it is documented in several books, that in every factory there is an invisible orhidden factory that keeps on producing losses. The reference is to the produce of non-conformities, that create extra non value addition work(s) resulting in huge loss of materialand allied resources consumed as also loss of production potential.

What is possibly not documented hitherto, is the fact that there is a CORNER OFEXCELLENCE in every factory or work place. It implies that every factory on some day, insome shift, on some machine, some worker(s) using some material, instrument(s) and alliedfacilities did produce a superb piece. The phrase CORNER OF EXCELLENCE in this contextmeans a stratum, from among the many strata possible by classifying the data by levels of theseveral factors listed in stanza 5.4 above, that gave rise to this exemplary performance. Now,the job of being perfect is simple, if only the data were so maintained that each output ofinterest can be traced to each input of concern, thoughtfully listed in the relevant cause andeffect diagram. The art of stratification will help in locating this stratum or corner. Theelements of this stratum should form the work standards. The entire, associated humanresource needs to be oriented to follow these standards. The produce will be non-conformityfree, a status superior to six sigma or ppm status being craved for ! What else do we want? Isit not an indispensable tool?

SCATTER DIAGRAM 45


6SCATTER DIAGRAM

6.1 DEFINITION

The dictionary meaning of the word �scatter� is sprinkle, dissemination or dispersion.In the present context:For quality parameter of interest, there is need to assess the nature and or the degree of

impact on the �quality parameter� of the resulting output as reflected by the observed changescaused by the corresponding variations in the �input parameter�. The former is called the�dependent� variable and the latter �independent variable�. Attempts are made to get the yieldor the behaviour of the dependent variable in desired range by monitoring the causing orindependent variable, in appropriately predetermined limits as indicated by the nature of therelationship between the two. The dependent variable is plotted along �Y-axis� and theindependent along �X-axis�. Such a diagram is called scatter diagram. See Figure F6.1.

Figure F6.1 A scatter diagram showing production of clinker in tons per workinghour versus draught before waste gas fan in a dry process cement plant.

0

10

15

20

25

30

35

Y

10 20 30 40 50 60X

Draught before waste gas fan

Appear to be freak observations

Pro

duct

ion

ofcl

inke

rin

tons

per

wor

king

hour

ofki

ln



6.2 UTILITY

It is used to indicate whether a pair of the parameters or the variables consideredtechnologically mutually interdependent, are really so in the process under examination. If so,the nature of dependence�linear or non-linear and the degree or strength of the relationship.

Thus, it is useful in identifying the vital few parameters that play a dominant role in thefinal value of the product parameter of interest. It also indicates the approximateadvantageous level of the process parameter to be aimed at, as also the direction for furtherattempts for achieving higher targets.

It also helps in substituting a cumbersome time consuming existing test procedure by analternative less time consuming or an expensive one by an alternative cheaper method, yetequally reliable one for speedier information necessary for feedback for effective control ofprocess and or product parameter. An example of this appears in Figure F6.2

Figure F6.2 Showing relationship between initial and final setting times of cement.

6.3 PRE-REQUISITES

In industrial situations, the pre-requisites for deriving maximum benefits from the use of thistechnique are:

Selection of the problem, keeping in mind the likely resulting potential benefits�direct,indirect, tangible or intangible or any combination of these.

Listing of all independent parameters that are expected to influence the dependentparameter of interest. It needs to be noted that assessment of relationship is not possiblefor the factors not included in the list.

Care should be taken that such factors are by and large measurable or at leastquantifiable and controllable. The unit of measurement should preferably be less than

40 60 80 100 120 140X

60

80

100

120

140

Y

Fin

alse

tting

time

(hou

rs)

Initial setting time (minutes)

SCATTER DIAGRAM 47


one-tenth of the observed range or the specified tolerance whichever is lower. Like wisethe repeatability and reproducibility errors together should not exceed one sixth of theobserved range or the specified tolerance.

6.4 PROCEDURE

The approach consists of collecting a set of 25 observations on corresponding dependent andindependent variables or parameters of factors of concern. Care needs to be taken that thedata are spread over a period of time that gives a fair chance of occurrence or representationto the factors that are beyond any feasibility of control. The data are then plotted on a graphto draw scatter diagrams, of dependent variable versus each of the independent variable forinterpretation. The scale for plotting the graph or the scatter diagram may be so chosen thatthe spread covers an area close to a square and large enough to be adequately visible.

A few vital, among these, that indicate relationship worthy of exploitation, are thenchosen for determining the ranges between which these should be controlled to get thedependent parameter or variable in the acceptable range. The control is exercised bymonitoring through samples of suitable size at appropriate intervals. If necessary, these aretaken up for further in depth study to assess multiple correlation and the likely impact withthe help of more advanced techniques.

6.5 COMMONLY OBSERVED SCATTER DIAGRAMS

The Figures F6.3 (a) to (d) are examples of commonly observed scatter diagrams that fail toprovide evidence of any worthwhile relationship. This is so, whenever, the points are soscattered that they are spread over a square, circle, ellipse or rectangle (vertical or horizontal).

Figure F6.3 Showing scatter diagrams indicating no worthwhile relationship.

X X

Y Y

0(a) (b)

X X

Y Y

(c) (d)



The Figures F6.4 (a) and (b) provide evidence of weak linear positive relationship. On thecontrary the Figures F6.4(c) and (d) indicate the presence of weak linear negative relationship.The relationship is termed linear if a line can be drawn through among these to fairlyrepresent the scatter, as shown in these figures. It is said to be positive if �y� is seen to increaseas �x� increases & vice versa and negative if �y� decreases as �x� increases and vice versa.

Figure F6.4 Showing scatter diagrams indicating weak linear relationships.

Figures 6.5 (a) and (b) are illustrations of strong linear negative and positive relationshipsrespectively. Figures F6.5 (c) and (d) demonstrate strong non-linear (curvilinear) relationshipswith peaks facing up and down respectively. The strength of the relationship is judged fromthe extent of the observed deviations of the scatter about the line or the curve representingthe same. The narrower or closer the spread of the scatter, the stronger is the relationship.The slope of the line or that of the curve in various segments provides a measure of the likelychange in the dependent variable caused by a unit change in independent variable. Therelationship is said to be perfect or ideal, if all the points lie exactly on a straight line or anyspecific curve, that is, with zero deviations about the line or the curve drawn, as the case maybe.

6.6 INTERPRETATION

The absence of any worthwhile evidence of relationship between the two parameters, when itis considered very likely from allied technological considerations, provides an important cluethat either the process lacks control or there are other pertinent factors which arecamouflaging this. These need to be investigated and taken care of. Once this is done, therelationship which was hidden or dormant hitherto, might surface up to yield the extrainformation being sought. This ought to be confirmed by collecting fresh data, that is by

X X

Y Y

0 0(a) (b)

X X

Y Y

0 0(c) (d)

SCATTER DIAGRAM 49


repeating the study by taking the extra indicated precautions. This assumes that the data aregenuine, reliable and have desired accuracy. False data can mislead with �disastrous�consequences.

Like wise an evidence of relationship should be viewed carefully, particularly when noneis expected. These are termed spurious or non-sense relationships or correlations. Suchsituations have been known to have led to abuse of this simple yet useful tool.

Apparent anomalies should be taken with a pinch of salt. These, however, serve a veryimportant purpose, by revealing the need to reconcile the diverging aspects�the views oropinions based on associated scientific or technical knowledge and the facts observed from thelive data. The attempts to reconcile this gap between the theory and the ground reality haveoften led to identification of contributory factors, hitherto unsuspected. The follow up leads tothe solution to the problem on hand.

A word of caution, the relationships are considered valid only for the range or the intervalof the variation observed. Interpolations or forecasts within the range experienced are valid.Any extrapolation is not permissible. If the technical considerations or the analysis of data orthe need justifies, the fresh data should be collected in the extended range of interest andanalysed to confirm its validity before implementing the recommendations resulting from theanalysis of data. It is a healthy practice to make confirmatory trials before implementing therecommendations of the findings on routine basis. Interpret the scatters intelligently.

6.7 DETECTION OF PRESENCE OR OTHERWISE OF RELATIONSHIP BETWEEN APAIR OF VARIABLES, FACTORS OR PARAMETERS

(a) Linear

The procedure described below is followed (See Figure F6.6)

X X

Y Y

0 0(c) (d)

X X

Y Y

0 0(a) (b)

Figure 6.5 Showing scatter diagrams indicating strong linear and curvilinear relationships.



Consider about 25 pairs of observations and plot the scatter diagram. There are 26 pointsin the said figure.

Draw a median parallel to the x-axis such that half the points are below the median lineand the other half above it.

Like wise draw a median line parallel to the y-axis such that half the points are to its leftand the other half to its right.

Notes:1. If there are a total of odd number of points or pairs of observations, say (2n+1), then

draw lines parallel to x-axis and y-axis through (n+1)th point, after these have beenarranged in non- descending order of y and x values respectively.

It may so happen that more than one point may lie on the median line.

2. If there are a total of even number of pair of observations or points, say (2m), then drawlines parallel to x and y axes such that these pass half way through (m)th and (m+1)thpoints after these are arranged in non-ascending order of x and y values respectively.

It may so happen that both (m)th and (m+1)th or more points may lie on the median.

3. This is easily done by running a scale parallel to x or y axis as the case may be, till thecriteria stated above are satisfied.

Designate the four quadrants thus formed as I, II, III and IV. Count the number ofpoints in these quadrants respectively, say n1, n2, n3 and n4. Ignore the points fallingon the median lines, if any. In Figure F6.6 n1 = 1, n2 = 12, n3 = 0, n4 = 11 . Two pointslie on the median parallel to y-axis

Note:In the absence of any relationship, hypothesis of assumption of n1 = n2 = n3 = n4 should

hold valid, except for random or chance deviations of no consequence. In particular, ideally(n1 + n3) should be equal to (n2 + n4). Statistically speaking, there should be no evidence ofthese differing significantly.

In the event of presence of any worthwhile linear relationship between the two, thedifference between (n1 + n3) and (n2 + n4) should be large enough, wider than what can begiven a benefit of doubt on account of random or chance fluctuations. For significance, underbinomial assumption either should exceed (n/2 + √n) at about 95 percent level of confidencewhere n = n1 + n2 + n3 + n4.

If (n1 + n3) is significantly greater than (n2 + n4), then the relationship is deemed positive.On the contrary if (n2 + n4) is greater than (n1 + n3), then the relationship is deemednegative.

In this case, see Figure F6.6,

(n1 + n3) = 1 and (n2 + n4) = 23;

n/2 + n = 13+ 4.9 = 17.9, 23 > 17.9

Thus there is very strong evidence that the �Tensile strength� is linearly related to �percentcarbon equivalent�, and the former can be achieved through control of the latter inappropriate range provided other factors are not disturbed and other conditions remain thesame. Also, this relationship can be used to predict Tensile Strength given the value ofPercent Carbon Equivalent.

SCATTER DIAGRAM 51


Figure F6.6 Detection of linear relationship between two variables, factors or parameters.

(b) Non-Linear

To verify presence or otherwise of �specific� non-linear relationship, standard transformations,available from published literature are made such that the transformed data exhibit linearrelationship, which in turn can be verified with the help of procedure described above.Standard statistical computer soft wares can also be used to obtain solutions to large or morecomplex problems.

6.8 EVALUATION OF LINEAR RELATIONSHIP AND COMPATIBLESPECIFICATIONS

(a) Determination of the equation of the line

Objective Method: Let there be �n� pairs of observations for �x� and �y�. Then, the regressionline is given by y = a + bx, where �b� is the regression coefficient and �a� the constant, calledthe intercept on y-axis.

The constants �a� and �b� are so determined that the sum of square of deviations betweenthe observed values and the corresponding estimated values, from the above equation orread from the relationship graph, is least. This is why, this line is called the line of �bestfit�. The differentiation procedure for minimizing, estimates the values of the constants �a�and �b� as under:

( )( ) ( )= / Σ Σ 2b x � x y � y x � x and a = y � bx

Alternative, simple procedure is given below: (see Figure F6.7)

3.50

14

16

18

20

22

24

26

28

Y

3.6 3.7 3.8X

IVIII

II I

Tens

ille

stre

ngth

(Kg/

mm

)2

Percent carbon equivalent



Divide the scatter of points into three segments by drawing two vertical lines parallel toy-axis, such that there are almost equal number of points in each of these segments A, Band C.

On the lines similar to the procedure, described in 6.7(a) earlier, to detect relationshipbetween two variables, draw median lines parallel to x and y axes for points in segmentsA and C. Let the points of intersection of these medians be designated P1 (X1, Y1) and P2(X2, Y2).

Join the points P1 and P2. The resulting line approximates the line of �best fit�, called theregression line. In actual practice, this is as good as the one derived by the objectivemethod. It is sometimes more realistic than the objective method because it takes care ofthe freak values, if any. Read the coordinates of the points P1 and P2. The equation of theline passing through these points is given by:

y = x (Y2 � Y1) / (X2 � X1) + Y1 � X1 (Y2 � Y1) / (X2 � X1)

In Figure F6.7 P1 (X1, Y1) = (3.59, 26.0) and

P2 (X2, Y2) = (3.73, 21.1)

Thus the equation of the line is y = � 35 x + 151.65

Figure F6.7 Determination of equation of the regression line.

(b) Estimation of the error

The estimate of the residual error (S0) is given by : S0 = √[S (y � y*)2 / (n�1)]

Where y and y* are the observed and estimated values of y respectively for given x.

The estimate of the residual error is useful in determining the specification limits of theindependent variable x for given specification limits of dependent variable y. To find these:

14

16

18

20

22

24

26

28

3.5 3.6 3.7 3.8

P (X , Y )2 2 2

P (X , Y )1 1 1

AY B C

Tens

ille

stre

ngth

(Kg/

mm

)2

X

Percent carbon equivalent

SCATTER DIAGRAM 53


draw lines parallel to the line of best fit at twice the distance of the estimated standarddeviation, for 95 percent confidence.

Given upper (U) and lower (L) specifications, of y, the corresponding specifications for x,that need to be met may be determined as shown in Figure F6.8. Of course if higher y ispreferable through further increase of x, keep on trying it, till it reaches the convex peak. Theadvantage is found in the relaxed control of x to get simultaneously y even in narrower limits.See the next stanza.

Figure F6.8 Method of arriving at the specifications of x, given those of y (Linear Relation)

However, if the relationship is non-linear, say parabolic (see Figure F6.9), the specificationof y should be chosen close to its peak, that is, the maximum value of y should correspond tothe point of intersection of the tangent with the y axis. For a given specified range of y, therange required for control of x, the corresponding process parameter, shall be the widestpossible under the circumstances and therefore least rigid control for greater economy.

The criteria may be carefully noted for the positive and negative relationships. The wrongpractices, generally in vogue out of ignorance, must be understood and avoided. It may also beunderstood that feasible limits for x are not possible if the vertical distance between thedotted parallel lines indicating natural bounds of variation around the regression line or pathor curve is greater than the specified tolerance�the interval between the specified upper andlower limits, that is if the Residual Variation of y is greater than the specified ToleranceRange of y. It must also be appreciated that the line or curve predicting x given y is not thesame as that for predicting y given x. The deviations for prediction d1 and d2 respectively arenot the same.

30 30

50 50

70 70

2.0 2.03.0 3.04.0 4.0X X

X X

U U

U

U U

U

U

Y Y

Y Y

Right way Wrong practice

L L

L

L L

LL



Figure F6.9 Establishing process tolerances (Curviliner Relation).

Product parameter: y, Process parameter: x

(c) Correlation Coefficient

Correlation coefficient is a measure of the degree of the strength of relationship between thetwo variables. It is generally denoted by �r�. Mathematically,

( )( ) ( ){ } ( ){ } Σ Σ Σ 222r = x � x y � y / x � x y � y

The calculated value of �r� can sometimes be very misleading. For example, seeFigures F6.10 (a, b, c, and d), wherein r = 0.83. Figure (a) reflects a situation commonlyexperienced in industry when two variables are fairly well related amenable to the proceduresand interpretation discussed herein. Figure (b), on the contrary, clearly indicates that therelationship is not linear and the optimal value is indicated by its peak. The linear relation inFigure (c) will undergo a major change, if the freak point shown in square is ignored. Theslope of the line will reduce and all the rest of the points are seen to lie perfectly on thestraight line. Thus, the corresponding change in y for unit change in x will be less.Simultaneously the value of �r� will change from 0.83 to 1.0, indicating a stronger relationship,strongest or perfect in this particular hypothetical example. In contrast, the linear relation inFigure (d) will vanish, if the freak point shown in the figure is ignored, implying that evenperfect control of x is unlikely to be of any help whatsoever in reducing the variation of y inthis particular hypothetical example, indicating the need to search for other factors, thatmight matter. It is not unlikely that the factors, least suspected on the basis of available

XLy

LxLx

Ly

Uy

Y

Ux

Uy

28

32

36

706866

28

32

36

Specified

Upper tolerance for YLower tolerance for YTolerance range for YDesiredUpper tolerance for XLower tolerance for XTolerance range for XIf X is temperature, it is economical to maintain it at lower level.Note: By exploiting peak region of parabola the tolerance of ‘X’ may be, relaxed(approximately doubled, from 0.35 to 0.75) with simulataneous reduction in variationof Y (by approximately half from 9.25 to 5.00) thus gain from this optimality may beconsidered to be four fold nay sixteen fold in terms of variance.

Example1

38.0033.00

5.00

68.3067.55

0.75

236.5027.25

9.25

69.00 or 67.0068.65 or 66.65

0.35

SCATTER DIAGRAM 55


technical knowledge, have been innocently ignored to such a large extent that (the processvariation of this parameter has increased substantially over what it used to be earlier) thesehave become important.

Figure F6.10 Hypothetical scatter diagrams that caution from pitfalls.

Such pitfalls can be considerably avoided by plotting a scatter diagram and looking at thepattern of the spread of the points with a special eye on freak points. In fact the proceduregiven for evaluation of linear relationship and compatible specifications for drawing aregression line, has a built in system, to ignore the misleading contribution from freak points.The pair of values corresponding to freak points, observed if any, should be ignored and �r�calculated afresh, from the rest of the homogenized data only. Thus, the scatter diagram is anindispensable tool, for valid inferences, whenever the relationship between a pair ofparameters merits consideration.

6.9 SUCCESS STORIES

(a) Jute Mill

The motors are used to run the looms. As the speed of the drive motors is increased, so is thenet speed of the looms or production increased. However, beyond certain stage there isslippage of belts and the net speed is adversely affected. A scatter diagram of net productionversus speed of motors was plotted. It resembled the shape shown in Figure F6.5(c). Theoptimal level corresponding to the peak prescribes the most advantageous speed at which themotors ought to be run.

0 0

0 0

4 4

4 4

8 8

8 8

12 12

12 12

16 16

16 16

20

20

4 4

4 4

8 8

8 8

12 12

12 12

Y Y

Y Y

X X

X X

(a) (b)

(c) (d)



Like wise, the optimal levels with respect to speed and breakage rate in spinning wasexamined. It needs to be noted that these optimal levels are very likely to differ from onefactory to another, depending on many factors such as raw material, environments andmaintenance of machines. Therefore, utmost care should be taken to assess the nature ofrelationship. It should be confirmed for its repeatability and reproducibility, beforeimplementing the same. In fact process data should be monitored to see that the currentoperations conform to the predicted band about the regression line. Deviations beyond theband, if any, should be looked into for avoidance, if the consequences are adverse and forretention, if these are favourable.

(b) Automobiles

The fuel consumption and speed of cars or automobiles in general bear relationship similar tothose shown in scatter diagram in Figure F6.5 (d). It is thus advantageous to run these atoptimal speeds corresponding to the lowest point on the curve for economy of the fuel, thereserves of which are limited. The other accompanying tangible benefits are longevity of thetyres, engine and body of the vehicle. The intangible benefits include improved safety andreduced pollution.

(c) Brewery

A brewery, interested in monitoring the recovery or the productivity from barley, hadelaborate system in place. It consisted of processing a sample of barley in controlled conditionsin the laboratory. The plant recovery percent was compared against the observed recoverypercent in the laboratory, instead of the ideal, one hundred percent, and rightly so, to allowfor the differences in intrinsic quality of the barley.

A scatter diagram of the plant recovery percent versus the laboratory recovery percentwas plotted. It resembled the Figure F6.3(c). It was very revealing and surprising to themanagement to know that:

the two recoveries did not bear any relation between themselves, and

the process variation was more than that observed in the laboratory, no second thought,it is just a known expected phenomena.

This lack of relationship, does not reveal (rather conceals) the factors for high variation inthe plant process recovery. Thus the objective for which elaborate laboratory was set up athuge cost and the introduction of test system appeared to have been defeated.

It pointed out the need for looking into the reasons for this among other associatedprocess factors, which are camouflaging the expected relationship, including accuracy ofdata�the method of estimating the plant recovery and testing in the laboratory encompassingthe use of chemicals (their concentration) and procedure (such as durations). May be, thestandard practices have been laid down, but are not being adhered to. If not, these need to bestandardized.

(d) Bicycle

It was observed during the assembly operations of the bicycle that the pedal was hitting thechain stay instead of having a free movement, even though only the accepted componentswere being assembled. The inspection department blamed the design department for its faultydesign. The design department in turn blamed the inspection department, which had taken

SCATTER DIAGRAM 57


the entire responsibility of assuring that the process conforms to the laid down proceduresand the product satisfies the specified dimensional tolerances. The data on the magnitude ofinterference and the dimensions of the suspected parameters of the components involved wereobtained and scatter diagram plotted. It resembled the Figure F6.3(a). The data were small,yet conclusive. The interpretation was explained to the design personnel; that even if all thecomponents were perfectly alike, the problem is likely to show up. They agreed to look, for theproblem, elsewhere.

The case examples discussed in sections (c) and (d) above affirm the utility of analysis offactual data (scatter diagram is, one such analysis), to think rationally to resolve the problemrather than engage in the game of blaming each other. The game of blaming vitiates the entireenvironment, detrimental even for the survival of the organization.

(e) Medical Research

Curative power of medicines for high blood pressure (B.P.) is being judged by relating theimprovement in B.P. over time after administering the prescribed dose. However, such studieswhen conducted along with control group, consisting of patients who have not been treated,and compared with the former did not provide good enough evidence of the efficacy of thetreatment. Thus there is need to conduct such valid statistical studies to avoid any hasty rawconclusions.

(f) Steam Boiler

Steam boilers are used to generate steam required for various processes in textile, vanaspati,pharmaceuticals, laundry attached to hospitals or hotels, rayon grade pulp, thermal powerhouses, starch and the like.

Common worthwhile pertinent relationships, that appear to be relevant for regularmonitoring, for control and improvement of productivity of the boiler house and the steamare:

The amount of the steam generated versus fuel consumed and

The amount of finished product or suitable weighted index of the product range versusthe amount of steam consumed

The above scatter diagrams were drawn for a pharmaceutical company. Once again thelikely relationship was not apparent. Generally, such anomalies are attributed to erraticvariation in the quality or ash content of the coal or the fuel received. The discussion amongthe personnel of the production and boiler house led to the following standard practices forcompliance:

Production Department will communicate their requirement of steam or lack of it to theboiler house a couple of hours in advance, to enable them to plan steam generation toavoid wastage of resources and

*Boiler house will arrange to feed coal in appropriate small broken sizes instead of hugelumps and at appropriate intervals to ensure proper combustion that in turn shall helpmaintain desired pressure and supply.

These measures reduced coal consumption for the same production by 35 percent.Financially, a saving of only one percent in a plant of normal size of sugar, textile orvanaspati will have the same implications.



Next, the relationship between the coal consumed and water consumed (used as anindirect measure of steam generated in the absence of suitable steam meter that will takecare of varying temperature and pressure) was established. Such a relationship could not beestablished in the new plant. Investigations revealed a fault in the design. The capacity ofgeneration was not compatible with the needs, leading to process failures causing productionjams. The design was appropriately modified.

(g) Shock absorbers of railway goods wagon

The shock absorbers consist of metallic discs embedded with circular rubber strips on eitherside (there are holes in the discs for proper grip of rubber strips on either side) placed in acylinder. These can be seen at the ends of the wagon, at platform level, parked at the railwaystations. There was substantial amount of rework for too porous or too hard rubber caused byless or excess rubber fed during assembly of discs before curing. The control practice in voguewas to cut strips as per standard length depending on the designed parameter of circles orrings on the discs. This apparently good practice was found to be unsatisfactory because thenet amount of rubber going into the process was not uniform, since rubber shrinks over timeand the interval between drawing of the strip and subsequent processes up to curing was notuniform. A relationship between length of the strip and its weight was established at thedrawing process stage. This relation was used to convert standards on length to weight. Thisexercise, would have been redundant, if the process sequence permitted cutting of the stripinto pieces of desired length at this stage and stored as such till next stage of assembly withdiscs. For ease of operation, go/no go limits were fixed for weight for various requirementsbased on the corresponding linear relation which resembled the shape shown in FigureF6.5(b). Both, the quality and productivity improved simultaneously.

(h) Vanaspati

One of the final parameters of hydrogenated oil monitored by the manufacturer is its finalcolour. It is likely to be influenced by the initial colour of the oil and the quantity of bleachingagent used as also its brand or quality. Reprocessing of batches to correct for non-conformityof colour or their diversion to different market segments was a routine phenomena. The ratioof quantities of bleaching agent used and oil was standard and constant over batches. Pastdata were used to draw scatter diagrams with final colour on y axis and initial colour on xaxis, separately for imported and indigenous bleaching agents also called activated carbons.Both these indicated fairly good linear relationships. The line for indigenous bleaching agenthad greater slope, indicating that less quantity of indigenous agent was required against theimported. However, the residual error for the indigenous variety was larger in relation to theimported, indicating the need for more homogeneous production of the bleaching agent by thelocal supplier. Had one common diagram been drawn for both the varieties, the possibility ofthe relationship being camouflaged or concealed is not ruled out. This, incidentally, is anexample of supplier wise stratification. Therefore, wherever appropriate and possible,stratified or group wise scatter diagrams should be drawn.

(i) Heavy electrical

A heavy electrical equipment plant manufacturer was facing the problem of maintainingdimensions of rotor slot trough, as per drawing. A list of the relevant influencing factors was

SCATTER DIAGRAM 59


prepared. These included process parameters and the dimensions of glass cloth and die. Thescatter diagrams of output versus input parameters were drawn one by one. This helped inidentifying the dominant parameters needing control.

This attempt also indicated when the die became due for recalibration for its dimension(s).Hitherto, in the absence of any knowledge, the die was unnecessarily tempered with, thoughin good faith, causing loss of effort, material and production.

(j) Chemical plant

A plant was manufacturing Titanium Dioxide. The final stage of process consisted of feedingslurry at the top of a rotary kiln and getting the finished product at the bottom. Theparameters of interest were quality and productivity of the output ; moisture and feed rate ofthe slurry; and temperature of the kiln in three zones.

Six scatter diagrams of Productivity per unit of time versus Quality; Moisture of Slurry;Feed Rate of Slurry; and Temperatures T1, T2 and T3 in the three zones of the kiln wereplotted. Adequate past data were used.

It was a pleasant surprise to learn that productivity and quality bore positive relation.Thus it was sufficient to study the behaviour of five input process parameters. Moisture andfeed rate did not seem to make any difference in the prevailing range of operationalvariations, implying that these were conforming to valid standards. There was strongrelationship with all the temperatures. The scatter diagrams among the temperatures, plottedsubsequently, taken two at a time, also showed positive relationships among themselves. Itwas decided to monitor temperature in the middle zone of the kiln only. The scatter indicatedreduction of the temperature by about 60 degrees for good results. The management includingtechnocrats were hesitant to accept the suggestion for fear of adverse consequences. Thepossibility of even an explosion in this delicate chemical process was not ruled out. They werepersuaded with the supporting evidence from the data that fears were not justified. Acompromise was, therefore, struck. It was agreed to exercise control of the temperature innarrower limits, through extra vigilance and simultaneously reduce temperature by 10degrees at a time and reach the target in stages, if no hurdles are encountered. This approachis synonym with evolutionary operations. The implementation of this strategy resulted inrecord production of better quality of Titanium Dioxide accompanied by reduced fuelconsumption. The tremendous gain in productivity of indirect costs can be imagined. Thatyear the production bonus too was highest ever.

The approach has been profitably applied in cement industry too, where the kiln operationsare similar.

(k) Rubber Manufacturers

A rubber manufacturer was rolling a tread to be shaped and cured into a cycle tyre. The crosssection of the tread, along its width, looked as shown in Figure F6.11. Its weight per meterwas an important parameter for control. The procedure consisted of cutting one meter lengthat regular intervals of time, checking its weight for conformance and making necessarycorrections in the process, if warranted.

The width, thickness, length and density of the rubber compound together determine theweight. The compound and its density were found to be fairly standard. The control of thewidth too was adequate. Samples of standard lengths were cut. Thus only thickness wasdesired to be investigated.



The thickness at three positions, one each at the ends and one in the middle along thewidth of the tread were associated parameters of concern. The process data were collected onthe samples being tested for weight (w) and on thickness at the three positions (t1, t2 and t3).Six Scatter diagrams were plotted�w versus t1, t2 and t3; t1 versus t2 and t3; and t2 versus t3.It was seen that t1, t2 and t3 were related among themselves. It was therefore possible toforecast weight (w) and exercise its control by monitoring thickness at any one of the threepositions. The repeatability and reproducibility errors of measurement of thickness of rubbersheet is more than that for solid sheets. Also, the more the thickness of the rubber sheet, themore is the error of measurement, inspite of the special flatted ends of the micrometersdesigned for use in rubber industry. Thus the revised system for control of the processconsisted of a thorough first off inspection to ensure parallelism of the rollers and the gapbetween the two and subsequent periodical check for thickness only at position t1.Occasionally position t2 was checked to re-affirm the parallelism. This procedure avoidedperiodic cutting of the tread and thereby the end waste during cutting of the tread into piecesof length equal to the periphery of the tyre prior to its insertion into moulds for curing. Thebonus was, the reduced handling.

(l) Conclusion

The foregoing case examples are not intended to contradict the technical knowledge available.These are intended to support; by highlighting each case on its merit; that there are somefactors, over which one has no control and there are others, with which one can play. It isnecessary to assess the impact of changes in the latter that will counter the adverse effects ofthe former to reap the accompanying benefits. As discussed earlier, the expected relations mayget dormant and invisible because of excessive variations. In such situations too, the support isprovided by bringing forth need to exercise better control by reducing variations in the factors.Scatter diagram is a tool to aid the technocrat to perform better.

Left side Right side

t1 t3

t2

Hump

Figure F6.11 Showing cross section of tread of cycle tyre.

HISTOGRAM 61


7HISTOGRAM

7.1 DEFINITION

French statistician A M Guerry is credited with development of histogram in 1833.A Histogram may be defined as graphical cum pictorial presentation of variation of a

single parameter observed in the data emerging from any process. Generally the horizontal x-axis represents the measurable value at appropriate scale and the vertical y-axis thefrequency with which the number of times the value has occurred in the form of a verticalrectangle as illustrated in the Figure F7.1.

Figure F7.1 Histogram of weight of tablets in grams.

7.2 THE ART OF MAKING IT

The measurements should be made preferably with least count approximating to one tenth ofthe specified tolerance or range of the observed spread over stable period of the process. In

3.53.43.33.23.13.02.92.80

5

10

15

20

25

30

Fre

quen

cy(n

umbe

r)

Y

X

Weiht in grams

Note : All items weighingbetween 2.75 to 2.85 gms arepresumed to have a weight of2.8 gms each and so on.



such cases, as mentioned above and as shown in Figure F7.1, the value is shown on x-axisand the frequency along y-axis in appropriate scales. This should do for most of the industrialsituations.

It is of course pertinent, that the data should be relevant, current, true, authentic andadequate.

7.3 UTILITY

It is good to remember that all data are subject to variation. It is necessary to understand thefactors that cause, the observed variations. The sources of variation are peculiar to eachprocess and hence the product. Different causes give rise to different patterns. It is helpful tounderstand their effect. We are then in a better position to conduct systematic investigations;to identify and trace the root causes of unwanted and avoidable variations ; and to search fortheir remedial measures. In addition to providing help in finding solutions to the problems, itcuts down on the time required to do so. The proposed remedies should be confirmed, suitablesystem evolved and implemented to sustain the gain with the help of run chart, the subjectmatter of the next chapter. Adequate education and training needs to be imparted to theworkforce to avoid hiccups. Simultaneous study of short term and long term variations isgenerally more informative, eye opener, educative and helpful in reducing and controlling thesame.

To take quantum advantage from studies speedily, it is extremely important to makehistograms for each process and or product parameter of concern, separately for each knownand controllable pertinent source of variation, such as machine, operator, material source(supplier or origin) and batch, shift, fixture. The histogram thus obtained should be comparedwith that expected, not necessarily symmetrical or normal. For example; run out, taper,ovality, centrality, eccentricity, life, number of or percent nonconformities in small sampleswill not exhibit normal shapes. Their shapes are peculiar to themselves and bear befittingnames.

These have established their utility. These are easy to make, convincingly communicativeeven to doubters. These are useful in breaking the proverbial ice and remove the hurdles inthe progress of finding solutions to the problems. At least, it provides clues for further probingultimately leading to solutions.

These have been innovatively used to provide crude yet good enough initial estimate ofprocess capability. Investigations aided by the clues provided by its shape have often led toimprovement and control of the processes resulting in enhanced productivity and quality ofgoods and services. Thus, does the society grow and prosper.

A glance at Figure F7.1 shows that the process average is around 3.1 grams and it variesfrom 2.75 to 3.55 grams. The shape is stable and symmetrical. The process capability may beestimated as: 3.55 � 2.75 = 0.80. It is considered excellent if the tolerance is 1.00, more so ifthe specifications are 3.10 ± 0.50; and poor if the tolerance is narrower than 0.80. On thecontrary, if the specified mean is different from 3.10, the situation can be easily remedied byappropriate action to get the desired mean. In this case, the feed rate of the mix may needadjustment.

These procedures, often over estimate the spread or under estimate the capability of theprocess. Appropriate methods to collect data supplemented by homogenization, that helps toalienate the contributions from assignable causes if any provide more realistic estimate of the

HISTOGRAM 63


process capability. We also, generally, under estimate our capabilities and potential. Thus, wemiss many wonderful platinum or diamond, life time opportunities.

7.4 PRE-REQUISITES

Before we proceed to discuss its applications, it would do well to recall that the quality hasbeen universally defined as fitness for use. The fitness needs to be translated into productspecifications encompassing physical, functional and reliability requirements. These in turn,require evaluation of compatible specifications of parameters of various inputs includingprocess(es). The specifications comprise optimal target and acceptable pattern of spread orvariation around it. The lesser the spread, the better it is.

As a first step, in any given situation, it is important to understand and list the factors orcauses of variation and divide these in to two distinct groups. One, over which one has nocontrol ; usually termed chance, common, natural or random causes. The other, which we canmanoeuvre, manipulate or control at will; usually termed assignable, special, unnatural, orsystematic causes. The amount of variation, observed when the process is under the influenceof former set of causes alone, is considered inevitable and is called process capability. It isestimated as six times sample standard deviation. Any increase in this is attributed to thelatter and is economically avoidable.

As a next step, the pattern of variation when the process is under the influence of chancecauses alone called state of statistical control, is familiarized. These are then compared withthose obtained from the live process data. Remedial measures are contemplated and executed,if the differences between the two are considered too wide to be palatable. Additionally, as apositive measure, studies are planned to improve quality through continuous efforts onreducing the variability.

Experience has shown that, often about 50 observations are considered adequate for rightinferences. However, one should be flexible depending on the situation. The considerationstake into account, desired accuracy, economics and feasibility. The latter in turn depends onthe rates of production and inspection, their costs, equipments and the skills required. Morethan 100 observations are mostly redundant.

7.5 SUCCESS STORIES

(a) Machine tools

A machine tools, engaged in job shop production, found conventional run or control chartimpracticable, since the production was over by the time meaningful feed back on conformityor otherwise of the dimensional parameters is received from the standards room responsiblefor measurement. A job shop is defined as a job of a few pieces at a time repeated rarely or ajob consisting of a large number of pieces but accomplishable in a short duration, say at mosta couple of hours. The job being referred here belongs to the latter category. Innovativeapplication consisted of setting up the machine and allied process parameters at satisfactorylevels and running the machine to produce just about 50 pieces. Histograms of all dimensionsof interest are made. If these looked like the one shown in Figure F7.1 and properly centeredaround mid value of the specified limits, the machine was run to complete the planned



production. The last about 50 pieces are collected separately and like wise checked forconformance with the help of histograms. If the pattern of the histograms both at thecommencement and completion of the job are acceptable, then the entire lot is accepted andtransferred to assembly. This assures right quality and simultaneously avoids one hundredpercent inspection. It is known, from the previous experience, that the machine runs stablyfor the duration of the job. Hourly or half hourly outputs, as deemed appropriate, may becollected in separate trays or bins to minimize the additional inspection work load that mightbecome necessary to segregate conforming and non-conforming pieces, in the event of anyunlikely failure of the system during the scheduled run. This is an innovative use of thehistogram for control of the process. Its usual role is that of a post mortem to verifyconformance to desired form. Even so, it provides invaluable guidance for immediate futureuse for improvement and control.

(b) Glass shells

Let us now look at the Figure F7.2. This summarizes the production per shift of 8 hoursduration, by a team of a dozen workers engaged in a particular synchronized sequence toproduce glass shells. The process is labour intensive. It is amazing that inspite of apparentlyidentical conditions the production has varied from a low of 340 to a high of 580 (about 80%more than the low value of 340), with an average of about 433. There are three peaks, oneeach at 370,410 and 490 besides a freak (yet feasible and repeatable performance level) pairof bars at 550 and 570 with average of 559. Imagine the gains of production, in the light ofthe fact that the raw material cost, which is directly proportional to the number produced, is

Figure F7.2 Production per shift of 8 hours of a particular model of glass shells.

333

Production (numbers)

344

57

18

26

37

Y

350 370 390 410 430 450 470 490 510 530 550 5700X

5

10

15

20

25

30

35

40

Num

ber

ofsh

ifts

Note : Shifts with productionbetween 340 to 360 areconsidered alike withproduction of 350 units eachand so on.

HISTOGRAM 65


only about one fourth of the total production cost. All other costs are fixed and per unit costcomes down with the increase of production. If production is enhanced from the present levelof say 430 to new level of say 555 (an increase of about 30 percent); the production costreduces to below 85 percent. It is considered possible if favourable levels of causative factorsthat prevailed during the period of freak beneficial period are traced and sustained. This callsfor study of histogram of various likely causative process parameters like temperature andviscosity of melt glass as also scatter diagrams of these factors versus production. Likelyrecurring gain of 15 percent in the cost of production sets guide lines for efforts andinvestment, in remedial measures. Slackness on the part of any one member of the 12member team or one�s absence even for short intervals has substantial adverse effect on theproduction of the team. This in turn hurts the interest of every member of the team in termsof loss of production incentive and also reduces the return to the company, besides likelyadverse impact on its competitive potential.

A provision for extra suitable person, capable in all operations, to replace the teammembers in rotation or in exigency, besides control of process parameters can completelyachieve the mission of augmenting production, productivity and profitability.

It is absolutely necessary to assess the potential gains versus extra cost involved, if any,in the remedial measures and confirmation of these through trial runs before permanentimplementation. Such reviews should be repeated for further identification and exploitation ofpotential for improvement. Similar situation was experienced in a large cycle manufacturingplant too.

(c) Chemical plant

Now look at four histograms shown in Figure F7.3. The difference in production of a chemicalbetween 239th and 240th days, as also between 258th and 259th days (variation between twoconsecutive days or short term variation) are negligible. The causes for these minor deviationsare difficult to locate for elimination. However, the collective pictures of 258th and 259th days

11 12 12 12 12

259258240239Day of the year

13 13 13 13X

Production

1

4 46

10

14

1820

19

Y

10

20

Num

ber

of h

ours

dur

ing

the

day

Figure F7.3 Histograms showing frequency of production of a chemicalper hour in 00’s of litres, for a pair of two consecutive days each.



together versus that of 239th and 240th days shows unambiguously a favourable change, adeviation observed over a longer period or interval of time called long term variation. Thepotential for increase of 1200 litres per day or about 4 percent is visible, once again at noextra cost except that of raw material. There is saving on energy, human resource and otherheads. The cause, for this beneficial change, is considered relatively easy to trace andperpetuate. Another aspect that can be noticed is that the production figures are recorded inmultiples of 100 litres, due to constraints of observing facilities. It will do well to record theproduction figures either, in smaller least count or by shift instead of by an hour to get betterpicture of variations. It is likely to be more useful in investigations.

Such situations are often encountered in manufacturing and service organizations andprovide enormous opportunity, ready to be harnessed.

(d) Heavy electrical

A heavy electrical equipment plant was engaged in drilling about 4000 holes for tubes incondenser plate, with tolerance of 40 to 52 in coded units. The percent non-conforming holesvaried from under one to almost hundred. Sometimes all over size, some other times all undersize. Some times even nil non-conformity too was observed. This fact by itself is indicative ofthe process being capable. The histogram, shown in the upper half of Figure F7.4 presentsfrequency of one hundred holes at random. This is suggestive of high average diameter and acapable process. Generally, the operators have tendency to play safe and set for under size toavoid over size holes, which are more difficult and expensive to rectify. The over size holesneed to be welded first with steel of the same alloy, before re-drilling to size.

We also observe, concentration of values at 45, 50, and 55 ; all multiples of five. Themeasuring instrument used had a least count of 5. This is too large in comparison with the

Figure F7.4 Showing histograms of diameter of holes for tubes in a condenser plate.

10

10

0

0

40

40

2019

6811 11

45 50 52 55

4

42 1

51USL Status

USL

LSL

Y

16

X

X

Beforecontrol

Aftercontrol

20

20

30

40

50

29

45

Fre

quen

cyF

requ

ency

Note : Any observation lying between 39.5 and 40.5 shall normally be recordedas 40. Fifty percent of these, though below 40, which perhaps should beconsidered non-conforming are invariably accepted.

18

HISTOGRAM 67


tolerance range of only 12. An instrument with least count of �1� was recommended. In themeantime, process control was initiated, with advice to the operator to make eye estimate ofthe size up to least count of one, even while using the same instrument, till new one isprocured. The histogram of one hundred holes of next plate is shown in the lower half of theFigure F7.4. It certainly shows distinct improvement. However, now bias in favour ofundersize is obvious and concentration of values at multiple of 5 persists, though to a lesserdegree.

(e) Plastic textile

In a plastic textile mill the diameter of wound bobbins was checked for conformance tospecifications with the help of go/no-go gauges. The number of non-conforming bobbins insamples of 20 bobbins each, varied from 2 to 12. Surprisingly, the number was always even,a very unusual and unexpected phenomena. It is very difficult to explain this unintended biason the part of all associated with the inspection and segregation. The differences between theobserved phenomena and expected natural phenomena (frequencies) are substantial. It givesundisputed evidence on the feasible scope of improving the validity of inspection results andhence a valuable support for subsequent efforts in improving conformity. See Figure F7.5.

Figure F7.5 Histogram (It is called Bar Chart, when applied to attribute data)showing frequency of non-conforming bobbins in samples of size 20 each.

(f) Tele printers

Now, let us proceed to look at Figure F7.6 showing histograms of inspection results of thesame 25 units on two sets of measuring electrical instruments, one each used by productionand inspection personnel. The glaring difference, between the two measuring instruments,arising from the bias, is beyond any shadow of doubt. Since, identity number of the unitsmeasured on each instrument had been maintained, the pattern of the scatter of theseobservations will reaffirm the conclusion arrived at.

Observed

Expected

Observed Expected

0 0 12 0 4 0 20 0 8 0 14 0 2 0

00

5

10

15

20

25

1 2 3 4 5 6 7 8 9 10 11 12 13X

0 1 2 4 6 9 10 8 7 6 3 2 1 1

Y

No.

of s

ampl

es

Number of non-conforming bobbins



The consequences of this bias, can be appreciated, if the acceptable ranges were 1 to 10 or6 to 15 or 11 to 20. Even though, the process is capable, the non-conformance might vary fromzero to one hundred percent, causing all kinds of chaos in the minds of all concerned and onthe shop floor too. The visible multi-peaks suggest that it is possible to produce components inthe narrower tolerance range of 5. To assess further scope, more evidence needs to begathered.

Unless, such behaviours are understood, the need for any remedial measure is never feltand hence the situation continues by default. These simple analytical reviews bring out thehidden truths, reveal the potential for improvement and provide vital clues to pursue for thesolutions. The efforts in these directions take the organization closer to the goals of ppm or sixsigma or zero non-conformity.

Figure F7.6 Showing histograms of measurements of a parameter (Amperes) of 25 componentsof a Tele printer, on two instruments, one each in use by production and inspection departments.

(g) Shaving blades

Let us look at Figure F7.7. It shows honing angle of 50 shaving blades. It is seen to varybetween 18 and 21 degrees. Since, any value between 17.5 to 18.5 degrees will be read andrecorded as 18 degrees and so on; the range of variation or dispersion may be considered to be(21.5 � 17.5) = 4.0 degrees, against the specification of (25.0 � 15.0) = 10.0 degrees. Processcapability index is defined as the ratio of the tolerance range to the corresponding processcapability. As stated earlier, the process capability is estimated as maximum variation (range)observed in a fairly large sample (50 is considered good enough for most practical situations)when the process is under the state of statistical control or the influence of chance causesonly. In this case the process capability is estimated as 10/4 = 2.5 or 250 percent. The processis really very good. It has the potential to surpass six sigma status if the settings affectinghoning angle are not ignored too much. The shifts, in the average, between 18 and 22 degrees,

X

X

5

5

Y

Y

Num

ber

ofun

itsN

umbe

rof

units

2 4 6 8 10 12 14 16 18 20 22Amperes (coded units)

Instrument in useby inspection

Instrument in useby production

HISTOGRAM 69


in such situations, shall be considered harmless by most in the industry. However, Dr. GTaguchi�s concept of loss function does not recognize this as an excellent situation. He definessocial loss equivalent to �k� times the square of the deviation from the target, the middle valueof the specification or the optimal value that delivers the highest performance. In this case,the target is 20 degrees. Assuming that the honing angle of 15 or 20 degrees renders theblade as scrap, the loss is its sale price, say Rupee one only. The present price of one blade isabout Rupees seven only. The deviation at this extreme is (25 � 20) = 5. Its square is 25. Thus25 k = 100 paise. Or k = 4. This implies that a blade with honing angle as good as 19.5 shallimply a social loss of 4 multiplied by the square of its deviation, which is equal to 20.0 � 19.5= 0.5. This equals 4 (0.25) = 1 paise each. Like wise one with honing angle 16 or 24 degreesshall imply a loss of 64 paise per blade, implying that the real worth of the blade to the userin terms of number of smooth shaves is only 36 paise. In the above example the honing angleis estimated as 19.5 degrees. It may noted that the distribution is symmetrical around itsmean and is considered satisfactory.

Figure F7.7 Histogram of honing angle of shaving blades.

Now let us consider another parameter of blade namely centrality. This has only upperspecification of 0.10. The histogram is shown in Figure F7.8. The shape of the distribution isnot expected to be symmetrical or normal. It is expected to be skewed to the right. About onefourth of the blades with centrality between 0.05 and 0.07 are considered avoidable throughprocess control. Thus the process seems to be capable of producing blades with centralitywithin 0.04 which is 40 percent of the maximum specified. This process is also very good. Theprocess capability index once again is 2.5.

It is worth recording that the production department wanted to import new capablemachines to manufacture blades to conform to desired specifications. New machines may beneeded for augmenting production, but certainly for not enhancing conformance. The presentmachines are in excellent condition.

In another factory, making automotive components, the production chief, made a similardemand for grinding machines. He was asked to assess the process capability with the aboveprocedure. He was assured that in the of event of these being found incapable, necessary fundsfor procuring new machines shall be provided, since the management is interested in marketingonly quality product. The study confirmed that the machines were capable and were in goodshape. The production chief satisfied himself that by using statistical process control charts,satisfactory product was obtainable and withdrew his demand for new machines.

X

20

16

12

8

4

Y

Fre

quen

cy

14 20

Honing angle15 25

LSL USL



(h) Die casting unit

Next, let us look at Figure F7.9. It represents a summary of 125 observations on hardness ofdie cast pieces of aluminium alloy, after annealing treatment. The men in the industry, whomatter, shall interpret 25 pieces concentrated on the left and another 50 on the border nearupper specified limit (USL) as acceptable giving a figure of 75 out of 125 or 60 percent asacceptable or conforming. Looking at it, from the view point of hitting the target,one mightsay that the conforming percent is zero. The diagram provides adequate evidence that theprocess is capable of maintaining hardness within a range of even half the specified range.The factors that need to be looked into are the annealing media, temperature, composition ofthe alloy, duration and the cooling rate apart from the heating system in the furnace and anysystematic differences among geographical locations in it. In the above case, it was mix up oftwo batches of different composition. The supplier was advised to avoid this. He complied withthe advice. His cooperation solved or eliminated the problem.

Figure F7.9 Histogram of hardness of aluminium alloy die cast component.

0.0 0.05 0.10X

Centrality

32

24

16

8

Fre

quen

cy

Y

USL

Figure 7.8 Histogram of Centrality of shaving blades.

74

14

7

12

31

USLLSL

Y

Fre

quen

cy

X

10

20

30

40

50

500 510 520

Hardness (codified units)

50

HISTOGRAM 71


A picture similar to Figure F7.9 emerged when resistance expressed in ohms of elements,called tracks used in the assembly of potentiometer was summarized in the form of histogram.

(i) Engineering�Automotive parts

One commonly comes across patterns of histograms similar to those shown for two batches Aand B in Figures F7.10. The reasons for these are traceable to biases among inspectors and ormeasuring devices. Besides, there is tendency to pass border line cases close to specifiedtolerance boundaries, even when there is abundant evidence that the processes are capableand these are avoidable through process control. Such components that are within permissibletolerance limits but do not conform to specified mean and normal pattern do create problemsduring assembly.

Figure F7.10 Histogram of a parameter of a component for two batches.

(j) Plastic fibre

Now let us look at a set of histograms of breaking load of plastic fibre, shown in Figure F7.11.The one at the bottom is the composite picture of production of a plastic fibre run on twomachines, extruders,in this case, A and B for two weeks I and II. It has four or more peaks.It is not uncommon to come across histograms which are flat called rectangular. These mayhave many peaks or modes. This is a clear sign of shifts in the process settings giving rise tochanges in the means. This becomes more clear, if we split the histogram into fourhistograms, one each for each machine and week. The highest peaks of these are respectivelyat break load of 500, 510, 540 and 580. Even, each one of these histograms has multiplepeaks. A look on the right side of top two histograms for extruder �A� shows gradual decline inits frequency. If this pattern were repeated on the left side, a crude estimate of processcapability may be assessed as 90. May be, further studies, shall provide a better (smaller ornarrower) estimate. In order to avoid non-conformities of breaking load less than 498 (or say,500); the process mean should be aimed at (500 + 90/2) = 545 or say 550. The parameters thatneed to be examined or taken care of, in this situation are temperature, speed and tension.

(k) Eelectroplating of automotive and allied components

The shapes of histograms might differ with the choice of grouping or the width of the classinterval, represented by a bar. Unintended wrong interpretations are possible for lack ofadequate knowledge and or experience. Figure F7.12 shows three histograms of same 54observations, made on the strength of etching bath in a plating shop (of rings used inassembly of automobile engines) with grouping (class width) of 4, 8 and 12 coded units

USLLSLUSL

A B Dimension

LSL

Fre

quen

cy

X

Y



respectively. The histogram shown at the bottom, clubs observations between 154 to 158, 158to 162 and so on. The middle histogram clubs the observations between 154 to 162, 162 to 170and so on. The histogram at the top, displays the same data by forming groups from 154 to166, 166 to 178 and so on. The histogram at the bottom shows largest number of peaks witha feeling of likely capable process. The one at the middle shows with not enough confidence oncapability of the process. While the one at the top, just shows a flat or a rectangular shape,indicating that the process has been left to itself with no central tendency and that theprocess variation is three times that of specified range or interval.

There is need to be cautious about divergent interpretations possible with different choicesof class interval. The choice of class interval should be judicious. Never the less, all the threeshapes do indicate potential for improvement. The thumb rule is that the number of classintervals should be equal to the square root of the number of observations available.

The tolerances for plated components of cycles are wider while that of printed circuitboards are narrower than the one's, just discussed. In all these cases, adequate attention is notpaid to maintain concentration of various baths within specified norms. Surprisingly, in allthe above three situations the observed variation is three fold of their respective specifications.Improved maintenance of the batches through guided replenishments and replacements hadsalutary effect on percent non-conformities. Simultaneous control of other bath parameters likecurrent levels brings further improvements.

Such rewarding results were also obtained in cell house, producing hydrogen and oxygen,in vanaspati manufacturing plants. Similar, was also the experience in the plating process ofwires used in the manufacture of telecommunication cables.

Extruder Week

I

II

I

II

I + II

10

20

20

12

20

40

60

0

0

0

0

0

A

A

B

B

A + B

Y

Fre

quen

cy

X405 425 475 525

LSL575 625

Figure F7.11 Histograms of breaking load of plastic fibrewith minimum breaking load of 498 in coded units.

HISTOGRAM 73


(l) Fertilisers

A most interesting fallacious use of histogram came to light in packing section of a largefertilizer plant. A histogram of weight of 200 bags picked up at random from a day�s output ofall machines showed a stable symmetrical shape like that in Figure F7.13. It was concludedthat the process was in a state of statistical control and the observed variation was more thantwice the specified tolerance range, some thing needs to be done to reduce the same. Thecause of excess variation, propounded by the technical experts was the presence of the dust inthe packing room interfering with the performance of sophisticated filling and packingmachines. Hence air-conditioning of the huge filling and packing hall was recommended, littlerealizing that the dust was emanating from the fertilizer itself. The ground reality of thesituation was that the settings or adjustments of various nozzles of different machines, werenot receiving due attention, for lack of conclusive information. The histograms drawnseparately for these sources confirmed the capability of the existing facilities. Air-conditioningwas avoided. Appropriate frequent dusting and adjustment of settings for standard weightyielded satisfactory results.

Y

10

10

10

0

0

0

11

12

4

11

11

2

2

20

655

1 1 1

8

(12)

(8)

(4)

10

5 56

10

10

10

9

Frq

uenc

y

X

2 2 2

7

4 4 4

230210190170150

Strength of etching bath (coded units)

Group(class interval)

Figure F7.12 Strength of etching bath, plating shop—Specifications 180 to 200 (coded units).



Figure F7.13 Histogram of weights of 200 bags of fertilizer (in Kilograms).

(m) Conclusion

Histogram is yet another simple tool that can be practiced, conveniently to review theperformance of a process for making improvements. All it requires is a pencil or a pen and apaper, in addition to existing facilities. It has one disadvantage. The summary of the datapresented in the form of the histogram ignores the sequence of production. Thus some vitalclues; of process imperfections, possible from the examination of the pattern of sequence; getlost. These get made up by the use of run charts, the subject matter of next chapter. The use ofhistograms compensates by highlighting, even small deviations from the target prominentlythrough review over convenient, logical and meaningful periods. The examples sited abovehave thrown open an enormous potential for its exploitation with benefits that one cannotafford to sacrifice.

The preceding, three tools, namely, stratification, scatter diagram and histogram play avital role in resolving the problems. These tools, provide very vital clues that help theinvestigations to trace the root cause as also their remedies. There might arise some situationswhen the problem does not get resolved and generate a feeling of failure. It would be unfair toaccept that status. The findings of the study, have taken us thus far, to reveal one way thatdoes not take us to our destination and provide leads and direction for our next steps to reachour goal.

12

8

15

0

5

10

15

20

25

30

1

3 3 34

9

13

17

20

22

25

2121

00 012

44.5

45.0

45.5

46.5

46.0

47.0

47.5

48.0

48.5

49.0

49.5

50.0

50.5

51.0

51.5

52.0

52.5

53.0

53.5

54.0

54.5

X

YF

requ

ency

Weight in kgs

RUN CHART 75


8RUN CHART

8.1 DEFINITION

A Run Chart may be defined as a graphical presentation of data on a chart with sequence ofproduction on the x-axis and the parameter of interest on the y-axis.

8.2 CHOICE OF SCALE

The time interval on x-axis may be as large a unit as �a million years� as applicable ingeological studies devoted to science of earth�s crust, its strata and their relations or changesor study of fossils�things preserved in strata of earth or anthropology, the science of man.Alternatively the unit may be a century, as commonly referred in history; or a decade as inthe case of census; or a year as in the case of usual annual financial reports of commercialorganizations. On the contrary, the unit may be as small as an hour or a minute or even asecond depending on the production rate or the speed with which the parameter underobservation undergoes a change of a magnitude that causes concern to the observer.

The scale on y axis should match the needs of visual appeal or perception for meaningfulinterpretation and convenient for plotting depending on the unit or the least count of theobservations available.

8.3 UTILITY

To make it a useful tool for understanding the bahaviour of the process, namely the resultingmean and the spread, it is desirable to indicate the desired target or the mean and thepermissible upper and or lower tolerances viz the acceptable boundaries.

The trueness and precision of the observations or the measurement should be goodenough for the purpose. The least count should be one tenth of the specified tolerance intervalor the natural spread of the process whichever is smaller. A smaller least count may add tothe cost while a larger may defeat the purpose. The repeatability and reproducibility errorsshould not exceed one sixth of the natural spread of the process or the tolerance interval forsimilar reasons.

Likewise, in the case of attribute data, expressed in numbers or percent should berecorded to the accuracy of approximately one tenth of acceptable or observed quality levelwhichever is better.



The interval for making observations, may be about half of the stable duration of theprocess; the spread, during which interval, is considered harmless. However, if the situationdemands as a part of special in depth study or otherwise, the observations on all itemsproduced may be plotted in the sequence of production, over the period of interest.

The success stories documented in (8.6) shall substantiate the benefits of satisfying theabove requirements.

8.4 PRE-REQUISITES

It is pertinent to emphasize even at the cost of repetition, that the data need to be relevant,authentic and adequate; save in exceptional cases when one may have to be contended with afew observations or necessarily deal with large data or whatever is available. The chart willbe able to portray only as much information as contained in the data.

Usually, about 25 to 30 observations are good enough to familiarize with the behaviour ofthe process. The period of observations should be wide enough to provide opportunity to allpossible chance causes or sources, over which one has no control, to have influenced theprocess. These charts help to identify the periods of stable and unstable behaviours of theprocess, as also working levels of means during these periods. Such periods are then used totrace the sources thereof, from the relevant process data, available from the documentedcheck sheets, specially designed to take into consideration the factors included in thecorresponding cause and effect diagram. The reasons for unacceptable pattern(s) ordeviation(s) need to be gone into and measures to avoid their recurrence are found andimplemented to make and sustain necessary improvements. The efficiency is judged by thespeed of detection of undesirable change(s), stoppage of process and its resumption to originalor desired status. The effectiveness is considerably enhanced if all the perceptible changesoccurring in the process are recorded faithfully as remarks or foot notes. The changes mayinclude details like change of operator, tool, inspector, gauge, voltage, die, jig, fixture,temperature, lubricant, viscosity. Also the details of the nature and quantum of correctiveactions taken, need to be documented like wise. These together help in developing andupdating standard work instructions that aim at minimizing the variability and sustaining it.

8.5 EXTRA BENEFIT

This is a very important and useful tool in sustaining gains actualized by identifying and solvinga quality problem for improvement with the help of the six tools discussed hitherto. It must,however be borne in mind that a judicious use of this tool can often result in break throughsolution to a chronic problem, even though it is primarily intended to solve sporadic ones.

8.6 HARNESSING CHARTS FOR IMPROVEMENT

8.6.1 Attribute

(a) Engineering industry�press shop�drawing operation

The process consists of drawing a circular disc, cut from a tin plate, into a hemisphericalbowl. This bowl forms a part of oil tank of an hurricane lantern. About ten percent of the

RUN CHART 77


pieces developed crack during the operation of the drawing. The cracked pieces werescrapped. A sample of ten consecutive operations was observed every hour and number ofcracked pieces was recorded. Twenty five such samples spread over three shifts wereobserved. As a thumb rule, the sample size, expected (average in the long run) to contain onenon-conforming unit each is considered appropriate and twenty five such samples at suitableintervals are considered adequate for the purpose. The data are shown in the form of a runchart in Figure F8.1.

Figure F8.1 Number of cracked pieces observed in 25 samples of ten each.

It is observed that:There are two runs of ten (one to ten) and eleven (twelve to twenty two) points (samples)

respectively, viz with one crack piece each only.Except for 11th and 23rd samples which had zero and two cracked pieces each, all the

remaining 23 samples had one cracked piece each. The knowledge of binomial probability tellsus that if samples of size ten each are picked up at random from a box containing largenumber of beads (90 percent green symbolizing good pieces and 10 percent red symbolizingcracked pieces) repeatedly twenty five times; then on an average in the long run; 9, 10, 5 and1 samples respectively will show 0, 1, 2 and more than 2 red beads respectively. The bars andthe curve, shown in the right portion of figure F8.1, compare the observed and expectedsituations.

Note:- A Chi-square test to examine the proximity or otherwise of observed and expectedfrequencies shows this departure to be much wider than that can be expected under similarsituations.

Both the above observed phenomena are considered non-random or un-natural. Thisimplies that there is strong possibility of some special or systematic cause behind thisconsistency (precisely, one piece, out of every ten consecutive pieces, produced in a small spanof time, getting cracked regularly), which is too good to be true. The visual display makes this,otherwise obvious inference, convincing.

Yet, most people, consider this situation as satisfactory or normal stable behaviour,expected of a process prone to ten percent non-conformance, out of ignorance. This createscomplacency and a golden opportunity, to look for improvement, is lost.

This led to observation of consecutive operations more minutely. It was observed thatwhenever a piece cracked, the operator would clean the die with a greasy cloth before re-starting. Subsequent to this, every 13th or 14th piece was getting cracked. This was observedthree times. To confirm this periodicity, the die was cleaned after every ten operations and lobehold, all the thirty pieces were found to be free of any crack whatsoever.

05 10 15 20 25 (Sample number)

1

2

3

Y

*

*

*

* * * * * * * * * * * * * * * * * * * * *

X

Frequency

Number of cracked pieces

Expected

Observed



A discussion with the chief engineer of the plant followed. It ran as follows:

What is (are) the cause(s) for the incidence of cracking?

The Indian supplier, of tin plates, does not take adequate care in packing and storing. Thetin plates accumulate a lot of dust and also become rusty. These cause the sheets to crack.

In support of his argument, he showed additional operations introduced by him, to removethe dust and the rust. He further added that the incidence was only one percent whenimported plates are used, without these additional processes. No doubt, had these extrameasures been not put in place, the incidence of cracking could have been much higher.

At this stage, the operation in the plant was visited. The two heaps, one of good pieces onthe back side of the machine and the other of cracked pieces on the right side of the operatorwere seen. By naked eye observation estimate, the proportion of rusty sheets visible to thenaked eye was about the same in both the heaps. By the logic propounded, the latter shouldcontain a higher proportion of rusty pieces. Thus there is anomaly that needs justification.

What possibly can be another cause?

The gauge variation in Indian tin plates is higher. The control of thickness of the plates isnot given due attention during rolling operation.

One piece each, from both the heaps, was checked. The observation did not corroboratethe above argument.

The chief engineer was again shown, the recurring behaviour of cracking when the die iscleaned after the incidence of cracking and its absence when the die is cleaned in anticipationafter every ten operations.

The above leads to the following conclusion:● Without contradicting the claims and the facts reported by the chief engineer, it needs

to be understood that an Indian tin plate may require cleaning of the die after everyten operations and dismantling and re-assembling of the die once a day for crack freeoperation.

● While processing imported tin plates, these intervals may be 100 operations and a weekrespectively.

Thus a status of ten percent non-conformance can and has been changed to a status ofparts per million (ppm) or six sigma status, yielding over ten percent additional production atno extra cost giving tremendous boost to profitability accruing from this operation.

By way of information, it may be added that:● The tin plates were in short supply and its distribution was regularized through quota

system at controlled price.● Extra over ten percent more production of hurricane lanterns was possible.● The scrap was in fact sold at higher price to manufacturers of products of smaller sizes.

What shall the scrap manager do?

(b) Engineering industry�Job shop�welding process

Number of non-conformities per ten meters of welding jobs, done during the day for agroup of eleven workers were recorded for two weeks - twelve working days. This number wasused as an index of performance of quality. The wisdom of clubbing variety of jobs performedby several workers is questionable. However it may be inevitable and not undesirable either;if it is a team work on huge jobs, the output of each worker is relatively small and theircapability on the variety of jobs is almost alike. Never the less, the index is monitored in theform of a run chart, shown in Figure F8.2. It sounds a signal if the performance is too good or

RUN CHART 79


too bad in relation to the past acceptable norms. It is seen that the incidence of 7.8nonconformities per 10 meters on the fifth day is conclusively not consistent with the averageof 2.5. This sporadic event signals adverse change and calls for search and implementation ofcorrective measures. It is unlikely to be worker dominant, since even the average of all theworkers has shot up. Its root cause is more likely to arise from system failure, such as; use ofwrong, badly preserved or inadequately heated electrode; low voltage of current supply. Thesystem belongs to the management.

Figure F8.2 Run chart showing welding nonconformities per ten meters per day.

The company luckily had the habit of recording findings of the inspection with pertinentdetails. These details were used to make a two way table called a matrix table, as shown inTable T8.1. This shows the nature or type of nonconformity and its likely source as hunchedby the professional inspector.

Table T8.1 Showing incidence of nonconformities stratified by its type and likely source

Source→ Current setting Electrode More gap Gas cold TotalType ↓

0 1 2 3 4 5

Under cut 10 0 5 0 15Porosity 8 4 0 2 14Blow holes 0 3 0 1 4Cracks 0 0 0 0 0Total 18 7 5 3 33

The nonconformities were also stratified by operators. It varied from ZERO per tenmeters for the best performance to EIGHT for the worst. The performance of the remainingnine operators using the same index were 0.5, 0.7, 1.0, 1.8, 2.1, 3.1, 5.2 and 6.2.

Equipped with these facts, a meeting was held with all the welding operators in thepresence of concerned technical support personnel. The status of 2.5 non-conformities per 10meters as also wide dispersions were brought to their knowledge to seek advice for

Y

0

1

2

3

4

5

6

7

8

9

X121110987654321

0 0

1.9 2.5

5.4

7.8

0.80.1

2.62.3

2.4 2.5

Day

Num

ber

of n

on-c

onfo

rmiti

es p

er 1

0 m

eter

s



improvement to reduce avoidable tangible and intangible losses. They expresseddissatisfaction with the current status and vouched potential for doing better. They wereencouraged to offer suggestions. They asked for bigger grinders and better maintenance ofwelding equipments, particularly their current setting. The former was not considered logicalbut conceded to give them a chance to demonstrate. The latter was accepted on the basis ofthe data of Table T8.1 and steps taken. It was however informed that inspite of poormaintenance, some of the operators were getting it right by starting work only after gettingthese set right. Such dedicated and committed operators were complimented and encouragedto train others.

At the end of two weeks, the performance was reviewed. It had resulted in breakthroughimprovement of lowering nonconformities to below 0.5 per 10 meters, one fifth of the pastfigure. Operators moral was high and they pledged to make it zero.

(c) Heavy engineering industry�fabrication

Fabrication is usually given indifferent attention at least during initial operations, becausethese jobs at this stage are neither considered fine nor precise enough. This was a virgin areafor application of process control. It was decided to make a beginning with two operators. Thefirst operation was cutting of steel plates with gas flame. The job was examined visually forits fitness. It was decided, in a meeting, to assess the finish quality by assigning anappropriate demerit score. Each imperfection should be classified into one of the threecategories, minor, major or serious and appropriate demerit score assigned to each plate thathas been cut.

A nonconformity that could be managed without any rework at the next operation wasdefined as minor category. Each minor deviation was to attract a demerit score of one.

A nonconformity that could be considered acceptable after rework at the next operationwas agreed to be classified as major category and assigned a demerit score of three.

A nonconformity that is not considered worthy of any repair whatsoever and is deemed asscrap was to be categorized a serious and assigned a demerit score of six.

Obviously a plate free of any nonconformity would attract a score of zero.A run chart for demerit score for each plate that is cut was initiated for two of the

operators designated A and B. The data thus obtained for four plates for operator A areshown at the bottom of Figure F8.3. However the demerit scores for both of them aregraphically presented in Figure F8.3.

The status was discussed among all the five operators of the work area in the presence oftheir supervisors and executives. All of them recognized the objective evaluation forcomparison of the performance offered by this procedure. They also visualized the scope forimprovement and control that the demerit run chart would provide. The system was extendedto all the five of them there after. The data for operator A and chart for all the five of themfor next three jobs each are shown in the same Figure F8.3. The performance that startedwith a demerit score of over thirty converged to a demerit score of under ten per plate. Withawareness aroused, it is considered possible to actualize further improvements. This gives asignal that any quality parameter that gets quantified becomes amenable to control.

It may be noted that serious nonconformity was never observed. It may be desirable toimprove upon the criteria of classification and scale of demerit score for the imperfections inthe gas cut plates, to make it more useful.

RUN CHART 81


0

5

10

15

20

25

30

35

40

45

50

Y

1

0

14

3

45

2

0

11

4

37

3

0

8

4

28

4

0

8

5

29

5

0

2

3

9

6

0

2

5

11

7

0

1

4

7

X

Dem

erit

scor

e pe

r pl

ate

Plate No.Operator A

Serious (6)

Major (3)

Minor (1)

Demeritscore

A

B

ED

C

Figure F8.3 Run chart of demerit score of nonconformities observed on a gas cut plate.

(d) Heavy electrical�condenser plate drilling of holes

A particular model of condenser plate was designed to have 3905 holes each. The specifiedlimits for the diameter of the circular hole were 30.40 + 0 .00 and 30.40 + 0.12 in coded units.After the job is done, all the holes are inspected with the help of go no go gauges of sizes30.40 and 30.52 respectively by the authorized competent inspector. He marks thenonconforming holes for repair. The under size holes are drilled to size, while the oversizeholes are first filled by welding with electrodes of same alloy and then re-machined. Therepaired holes are inspected again for their conformance.

The rework in mending undersize and oversize holes was substantial and unpredictable,sometimes too little and on some other occasions too high. These caused hurtful delays andescalated costs. The past data for twenty one condenser plates was available. These aretabulated in Table T8.2

The same data are presented in the form of a run-chart in Figure F8.4.A look at the above data shows that the number of nonconforming under size holes varied

from zero to 3905 while that of over size from zero to 24 only. It goes to show that operatorsare cautious to avoid the latter because it entails more resources and operations. One plateconsists of as many as 3905 or nearly 4000 holes. Another has 2344 nonconforming holes.Excluding these two the average number of nonconforming holes per plate are only about 12.4or 0.32 percent only. With as little as only one nonconforming hole in 2nd and 6th plates outof about 4000 holes each, one need have no doubt about the capability of the process to yieldzero nonconformity. The problem addressed with the aid of histogram has been discussedunder the caption success stories in 7.5 (d) of chapter VII. Two separate run charts one eachfor under size and over size holes are likely to be more useful.



Table T8.2 Number of nonconforming holes in condenser tube plateTotal number of holes per plate = 3905

Number of nonconforming holes

Packet number Under size Over size Total

0 1 2 3

1 2343 1 23442 0 1 13 13 0 134 5 0 55 13 0 136 1 0 17 3 0 38 5 0 59 3905 0 3905

10 6 17 2311 13 24 3712 8 0 813 2 0 214 3 0 315 0 21 2116 4 0 417 6 4 1018 26 0 2619 20 0 2020 24 0 2421 16 0 16

Overall (Average) 305.52 3.24 308.76Excluding 1st and 9th plates. 8.84 3.53 12.37

X0

5

10

15

20

25

30

35

40

Tota

l num

ber

of n

on-c

onfo

rmin

g ho

les

Y

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

3905

Serial number of condenser plate

Figure F8.4 Run chart showing number of nonconforming holes in condenser plate.

RUN CHART 83


(e) Engineering industry�job shop�machining operations

The shop was manufacturing turbine blades. The percent nonconformities was on thehigher side of 30. This contributed substantially to rework in many forms, scrap and delays.A run chart on percent nonconformities was initiated and all pertinent details, such asmachine or operation, nature of nonconformity and its likely source, were recorded. SeeFigure F8.5.

The number of components produced was small. Each component required machining ofseveral dimensions and profiles. Several workers worked on several machines. The totalnumber of nonconformities observed on all components produced during the day was divided bythe total number of parameters checked on all the components, and the ratio multiplied by 100,to arrive at the desired figure of percent nonconformities. This clubbing looks irrational, buttheir was no alternative. The number of items, required to be processed, was small. The jobschanged very frequently. Plus point was that, a couple of in depth studies, wherever possiblehad demonstrated, the adequacy of the process capabilities. This encouraged a look at thecumulative picture to look for the causes of nonconformities and their preventive measures.

For quite sometime the status did not change for better. The actions as indicated by thenature of nonconformities was not taken for fear of hurting production.

The matter was discussed with the executives responsible for planning of allocation of jobsto machines and monitoring its progress. They were equally responsible for quality, but it wasnot given adequate attention. The reason was the delayed information on the previous day�sperformance. It reached them after the production for the day had commenced and anycorrective measures will disturb the tempo of the production. The situation was remedied bysummarizing and presenting the status, as obtained till about an hour before the close of theday. It is logical to assume that the on going processes will continue to behave in the samemanner as hitherto. This enabled the executives to plan and implement the remedialmeasures before commencement of the production next day. The incidence of nonconformitiesdeclined to below five percent by the end of the first week and below one percent during thelast week of the month.

The accompanying tangible and intangible benefits were tremendous. The absoluteproduction went up three folds. With reduced non-conformities, the production of conformingproduct shot up four times. The delays got eliminated and deliveries advanced. It reducedcontroversies among human resource and enhanced mutual understanding to create healthycongenial environment worthy of emulation by others. The financial implications surpassed allexpectations.

The above applications demonstrate the role of systematic studies in identifying thedominance of weak areas and also, whether these pertain to the domain of workers ormanagement. The approach brings the two together in search of the solution to the problem onhand, rather than laying the blame on each other's doors. This cultivates harmoniousenvironments and makes the target of ZERO NONCONFORMITY assailable.

8.6.2 Variable

Quality is universally acclaimed and accepted as fitness for use. The degree of Fitness for useas perceived and demanded by the user keeps on moving upwards, thus necessitatingcontinual improvement. Absence of improvement is fatal.

Translating fitness for use to optimal product and process specifications entails detailedstudy, of all inputs; based on analysis of appropriate, adequate and true data.



Figure F8.5 Master Control System (CHART + DATA SHEET)—Attribute.

3

3

0

0

0

0

0.4

0

0

0

0

0

10

50

17

0

2

2

0

0

0

0

0

0.0

0

0

0

28

23

12

0

0

0

0

0

0

0

0

0

0

0

0

-

-

-

34

0

0

0

10

40

0

4

3.8

0

0.4

0

13-15

34

18

26

0

0

3.0

0

0

26

0

0

0

0

0

13-15

34

18A

104

99

0

0

0

0

15.0

0

44

40

0

4

0

0

4.2

0

0

0

0

0

13-15

35,3650

18

0

0

0

5

128

84

16-189

0

2

35,49

10

0

15.8

0

1,112

0

0

1.7

52

166

3

0

0

11-12

45,46

3,8

0

167

0

2

25.0

0

0.3

0

13-15

49,50

1,10

172

0

2

23.1

0

0.3

75

31

2

0

0

66

15

86

2

0

10-12

36,48

1,2

0

216

0

6

30.8

0

0.9

115

Passed

Operator

Stage(s)

Tool-fixture-jig

Operation(s)

Rework

Cutter

Machine(s)

Machine

Scrap

InspectionPrevious stage

No.

No.

No.

%

%

%

Classification by degree/ critcality of deviation:

Source wise detail-number only:

DATA SHEET

Source wise detail-number only:

*Repeaters are indicative of areas needing priotity attention.

X29

8463

0.4

28

2402

0.8

27

7620

0.0

10

105044

4.2

9

88126

3.0

8

104044

4.2

5

695104

15.0

4

809138

17.1

3

667169

25.3

2

743174

23.4

1

700222

31.7

Day:

Data sheet:Production No.Non- No.Conformities %

0

5

10

15

20

25

30

35

YPercent nonconformities observed in turbine blades

CHART

% N

onco

nfor

miti

es

RUN CHART 85


For instance, consider fitness for use of detergent for washing garments. It implies itspower to brightly clean the fabric at competitive price, preserve its colour and life, without inany way harming the hands or skin of the user on contact and emitting obnoxious gases thatmay be inhaled inadvertently. The user will indeed get delighted if in addition it benefits theskin and the gases emitted are not only pleasant but also useful to the body. It is, obligatoryto: thoroughly examine all the properties of all ingredients used; determine their optimal mixand the formulation procedure including preservation and packing to ensure shelf life againstenvironmental hazards. The optimal washing procedure also needs to be determined andcommunicated.

Consider yet an other example of shaving blades. The user demands more number ofsmooth shaves per Rupee. This needs to be translated into specifications for : the direct andindirect raw materials; equipment and facilities for production and testing; procedures formanufacturing, preservation, packing, shaving and disposal; finished product anddocumentation. Each of these needs to be dealt with in perfect and sufficient detail. Forexample, these shall include:

* Steel alloy, thickness, width, and strength.* Sampling method, size, frequency and acceptance criteria.* Capable equipments and their maintenance schedule.* Competent human resource.* Centrality and Honing angle of the finished blade.

The above list is not exhaustive!In all such cases, the optimal level of each process and product parameter is unique,

designated as the target. The ideal quality is achieved if each parameter is dot on the target.Any deviation from this target implies less degree of perfection necessary for fitness for use.This imperfection is quantified as social loss deemed proportional to the square of thedeviation. It may be recalled that Mean Square Deviation is called variance and its SquareRoot as standard deviation, universally used as a measure of variation. Emphasis oncontinual war on reduction of social loss to society by Dr. Genichi Taguchi is synonym withthe optimal approach to hit the target with minimal variation around it.

In some situations the loss or hurt caused by poor quality shall be many times more thanthe cost of the product. It might even prove fatal. Therefore such deviations need to be peggedat unavoidable minimum level in the situation on hand. This minimum keeps movingdownwards with Cost Effective�Improvements in Process Knowledge and allied Technology.The art, of translating customer's perception of fitness for use into compatible specificationsfor appropriate product and process parameters and taking all actions necessary to achievethese, is indeed putting quality function in place. This is popularly known as Quality FunctionDeployment (QFD)

Often, there is satisfaction in meeting the specified tolerances in legal sense. This resultsin unintended complacency and sustains the culture of stagnancy with associated disastrousconsequences. The leaders however do not remain contended and break the ice by makingimprovements in product and process designs through updating skills of human resource andallied systems with emphasis on augmenting return on investment (R 0 I). Dr. J M Juran, theworld quality guru, states the axiom that all improvements are project by project only. Thesure way to assure quality is indeed to build it into the product. All one needs to do this, inmost economical way, is to learn the right way and to do it right first time.

Let us see How charts help us in this noble task?



Prevalent Situation:In most industries even today, one product parameter of an item is viewed at anyone

point of time. Its literal or legal compliance to specifications including border line cases isinterpreted as satisfactory situation. In good faith, the human resource complacently sustainsthis culture. The data thus generated form a part of history and adore the archive.

Desired Scenario:It is desired to demonstrate that the following steps pave the way for alternate cycles of

improvement (breakthrough) and control.

* Present the data in simple pictorial form consisting of run or appropriate control chart.[See IS : 397 parts 0, 1, 2, 3 and 4]

* Supplement the above by a histogram. [See IS:7200 parts 1, 2 and 3]* Look for the short and long term patterns and trends of variation.* Use the clues provided by these events as guide to investigate the sources of variation.

Cause and Effect Diagram and associated Critical Failure Mode and Effect Analysisarrived at a brain storming session among all concerned catalyse and sharpen this activity.

* Confirm the findings.* Develop procedures to hit the target and to reduce the process variation around it.* Implement and monitor the procedure so developed.

All these steps require PASSION for improvement and little investment, if any, tochannelise the energies in the RIGHT path to become leader and reap accompanying benefits.Are there any options? Let us consider four examples. These represent Chemicals, HeavyElectrical and Automotive Components organisations.

(a) Fineness of a Chemical

In a chemical plant manufacturing fertiliser, pulverised rock fineness passing through 100mesh screen, is an important parameter. The practice was to look at, one observation at atime, at the conclusion of a batch. This does not communicate anything except that by andlarge the attribute or the parameter conforms or does not conform to the specifiedtolerance(s). A glance at all the observations compiled in column 2 of Table T8.3 for afortnight tells us about its Range, 4.0, the difference between the largest value (94.2) and thesmallest value (90.2) encountered. All this sounds satisfactory in the light of specifiedminimum value of 90.0 and 'no' alarming signals are perceived to act for improvement. Thesedata have been plotted in the form of a Run-Chart in the top of Figure F8.6. A careful looknow highlights trends, sudden jumps, stable and unstable periods, leading to provocation thatif reasons for such deviations or behaviour arising from avoidable sources are traced andacted upon, the product will be more uniform and hence superior resulting in better customersatisfaction and associated rewards.

RUN CHART 87


Batch Batch

S. S.

No. No.

0

1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

0

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.41.

MFG. MFG.

Date Date

Mar. Mar.

26

1

27

28

29

1

16

18

19

20

22

23

24

25 30

X X

Fineness Fineness

2

92.6

92.6

92.4

93.2

93.0

92.6

90.5

92.3

92.8

93.0

91.3

92.4

90.2

92.6

92.8

93.2

92.5

91.1

93.1

93.2

2

92.7

90.4

94.2

93.6

91.8

92.8

93.0

94.0

93.3

92.6

92.80

92.80

92.60

92.50

92.60

94.00

91.70

93.40

92.60

92.8093.60

R R

Moving Moving

Range Range

3

-

0.2

0.2

0.8

0.2

0.4

2.1

1.8

0.5

0.2

1.7

2.1

2.2

2.4

0.2

0.4

0.7

1.4

2.0

0.1

3

0.5

2.3

3.8

0.6

1.8

1.0

0.2

1.0

0.7

0.7

0.2

0.0

0.2

0.1

0.1

0.4

1.3

1.7

0.8

0.20.8

Table T8.3 Pulverised Rock Fineness Passing Through 100 Mesh Screen.

9090.5

9191.5

9292.5

9393.5

9494.5

95

Y

X

X0

1

2

3

4

5R-the moving range

X-the observed value

Figure F8.6 Pulverised Rock fineness Passing Through 100 Mesh Screen.



X

X

R

X

40

40

35

35

30

30

25

25

20

20

15

15

10

10

5

5Y

Y

0

90.0

1.0

91.0

2.0

92.0

3.0

93.0

4.0

94.0

R (moving) chart

X chart

0.85

92.6

Figure F8.8 Figure F8.7 + Control Limits.

Figure F8.7 Pulverised rock fineness (passing through 100 mesh screen).

9090.5

9191.5

9292.5

9393.5

9494.5

95

Y

X

X

R = 0.85

X = 92.57

0

1

2

3

4

5R-the moving range

X-the observed value

LCL = X – 3X

UCL = X + 3X

UCL = D RR 4

= 90.30

= 94.90

Rd2

Rd2

76

= 2.80

RUN CHART 89


APPENDIX: A8.1

Σx = 3702.8, n = 40, x = 92.57 = 92.6 Approx.

ΣR = 37.0, R = 0.925 = 0.92 Approx., D4 = 3.267, UCL = D4 R = 3.04, Note: LCL = 0Range at serial no. 23 is 3.8. It exceeds the upper limit of 3.04. Ignoring this, the revised

R = 33.2 / 39 = 0.85, UCL = 2.8 This time, none of the Ranges exceeds the upper limit.The ranges are thus homogeneous.

Process s = 2/R d = 0.85 / 1.128 = 0.76, (d2 = 1.128), 3s = 2.3, PC = 6s = 4.6

UCL x = x + 3s = 92.6 + 2.3 = 94.9, LCL x = x - 3s = 92.6 � 2.3 = 90.3, LSL = 90Desired x = 90 + 3s = 92.3 for Cpk = 1 or 100%, Non-Conformance = 0.135 %If x = 90 + 4s = 90 + 3.0 = 93 when Cpk = 1.33, Non-Conformance = 32 ppmPerformance During GOOD PERIOD is considered achievable and needs to be sustained.

R = 2.4 / 8 = 0.3, D4 R = 0.3 ( 3.267 ) = 1.0, s = 2/R d = 0.3 /1.128 = 0.27, PC = 6s = 1.62;For Cpk = 1.33, x = 90 + 4s = 91Economy of x = 91 versus 93 needs to be examined by comparing the extra cost and the

likely rewards both tangible and intangible.current Cpk = ( x � LSL ) / 3s = (92.5 � 90.0) / 2.3 = 1.1 or 110%.Assuming process to be in state of statistical control. It may be noted that it is not so.N.B. During 28 to 30 Jan, fineness was unnecessarily recorded up to 2nd decimal place�

of course the digit being zero always.If x = 91, Cpk = (91.0 � 90.0) / 0.81 = 1.23 or 123% (using the achievable value of s = 0.27)

x = 92.5, Cpk = (92.5 � 90.0) / 0.81 = 3.09 or 309% (current average)x = 93.0, Cpk = (93.0 � 90.0) / 0.81 = 3.70 or 370% (average maintained occasionally)

Figure F8.9 Figure F8.8 + Histograms.

9090.5

9191.5

9292.5

9393.5

9494.5

95

Y

0

1

2

3

4

5R-the moving range

X-the observed value UCLX

LCLX

UCLR

R

77

.............

........

....

....

....

.

.

.

.

..

..

..

.............

.......

....

..

.

Fre

quen

cy o

f XF

requ

ency

of R



This Chart is now supplemented by a Run-Chart of Moving Ranges (the differencebetween pair of consecutive observations) at the bottom of Figure F8.6. A look at this endorsesour first hand perceptions. In Figure F8.7, the averages (means) of the observed parameterand the moving ranges are also indicated. Now the timings of the sudden changes and thedegree of the changes are perceived with greater exactness.

In Figure F8.8, the control limits (For calculations refer IS: 397 Part 1 or the AppendixA8.1) have been superimposed. It further enhances the grasp of the knowledge about theprocess behaviour and aids more objective inferences. The interpretations are strengthened bycritical examination of the Histograms shown at the right end of Figure F8.9.

The 13th point on the �x� Chart and 22nd point on the �R� Chart clearly appear as extremeor un-natural occurrences. The period from 30 to 36 points both inclusive on �x� Chart andperiod from 28 to 35 points both inclusive on the Moving Range Chart look natural and verystable. It is necessary at this stage to pursue the clues provided by the time of occurrences totrace the causes that matter from the knowledge of relevant Cause & Effect Diagram andtraceable documented process records for taking appropriate action, to reduce the variationand to stabilise at the target or the level of one's choice.

This aspect is illustrated in detail in the next example.

(b) KR Dimensions at Positions R & S.

This is a typical case from a giant heavy Electrical Equipment Manufacturing Unit wherebatch size rarely exceeds a dozen. Figure F8.10 displays Run-Chart of 11 Turbine Blades inproduction sequence with dimensions at positions R & S designed to be identical to conform tospecified tolerances of 22.61 + 0.10. It is seen that none of the observations exceeds the upperlimit of 22.71 mm or falls short of the lower limit of 22.61 mm and hence the complacencythat all is �0 K� if not fine.

It is not realised that, glass of milk specified to be on the table, is not consideredsatisfactory unless placed at the centre or at least at a safe distances from its edges, if thetable happens to be large enough.

Figure F8.10 Run Chart.

The observation at position �R� of 10th blade dot on upper specified value does notnormally cause concern even though it is as serious a matter as the glass of milk at the edgeof the table in real life situation, if not more.

A look at the Run-Chart Figure F8.10, reveals that:* the observation at position �R� is always more than that at position �S�, though both are

designed to be identical.

22.61

22.66

22.71 USL

LSL

Position: RS RS RS RS RS RS RS RS RS RS RS

Blade No. 1 2 3 4 5 6 7 8 9 10 11 X

Y

RUN CHART 91


* the differences between positions �R� and �S� vary from 0.03 mm to 0.05 mm. Thismakes up 50 percent of the permissible tolerance of 0.10 mm.

* the observations vary from 22.63 mm to 22.71 indicating a bit of higher process averagethan desired.

All these symptoms indicate presence of systematic or non-random sources of errorstraceable to causes that are considered assignable or special or economically avoidablethrough appropriate remedial measures. This in fact reveals that potential for reducing erroror variation in the process and hence improvement. A brain storming session among CrossFunctional Team or group of members resulted in adequate work instructions shown at serialnumbers 1 and 2 of column 1 of Table T8.4, from the knowledge of associated causes andeffects. This may be considered as a simple version of Critical Failure Mode and EffectAnalysis (CFM and EA).

Table T8.4 Work Instructions

Symptom Cause(s) Remedy

1 2 3

1. Dimension at position �R� 1. Taper in the fixture 1. Check taper with the feeler gaugealways higher than that 2. Incorrect cylindrical and assure the use of correct fixture.at position �S�. profile of the cutter. 2. Check cylindrical profile and sharpen

if required before use.2. Incorrect level of setting. 1. Operator's strategy 1. Operator educated on capability and

to play safe to avoid futility/loss of playing too safe as alsoundersize/scrap. the benefits of setting on target.

3. Large variation between 1. Excessive run in 1. Check roller run out and correct ifblades. roller. needed.

4. Sudden shifts in 1. Presence of chips in 1. Operator educated on the rewardsdimension. the job. of a little extra care to keep chips

away versus prohibitive consequences.

All these measures were incorporated for the next job that followed. It may be noted thatthe specifications for the same operation for the new job for the same parameter are different;different target but same tolerance band. Tool or Fixture kit is not entirely the same. Theresults are shown in Figure F8.11. It is seen that the differences between the positions at �R�and �S� are almost eliminated, however the differences among the blades have increased,though yet not violating the specified boundaries. A repeat brain storming session culminatedin updating of above Table T8.4 with additions at serial numbers 3 and 4.


25.19

25.24

25.29 USL

LSL


Blade No. 1 2 3 4 5 6 7 8 9 10 11 X

Y



The implementation of these measures for the next job resulted in blades withcharacteristic as shown in Figure F8.12.

Fig. F8.12 Run Chart.

The Quality has tremendously improved. The average is very very (almost perfectly) closeto the mean of permissible extremes. Also the spread around this mean is almost one-fourthif not one-fifth of that experienced hitherto. This implies five fold reduction in spread.Accepting concept of loss being proportional to the square of the deviation viz. Variance, theimprovement is really worth 25 folds. Reduction in standard deviation to one-fifth isequivalent to reduction of variance to one twenty-fifth.

In fact variation has reduced to a level that the least count needs to be reduced and or theprecision of measurement needs to be improved upon for effective control at the level achievedand initiating attempt for further improvement in future.

It is this sustained approach of pursuing clues to symptoms, symptoms to causes, andthen on to sources of variation through Cause and Effect Diagram and associated CriticalFailure Mode and Effect Analysis for perfection of specific guide lines to concerned humanresource to facilitate production of product conforming to specified mean with least possiblevariation around it, that will result in alternative cycles of improvement and control. Once,these conditions are satisfied, the subsequent assembly operations will have none or leastpossible hazards.

(c) Automotive Components

This example differs from the previous two in the sense that data collection were pre-planned to examine the amount of variation contributed by various sources contemplated inthe Cause & Effect Diagram. The balanced data facilitates, valid assessment of thecontribution from each source. The parameter under study is the close gap of rings that formpart of engine assembly along with piston, from an Automobile Component ManufacturingUnit. The sources considered were; Mandrels (4), Segments of a Mandrel called packets (3 oneach mandrel), Positions within a Packet (3-Beginning, Middle and End) and Periods (3-consisting of one cycle each). The study was confined to a specific Model, Machine andOperator. At each position of a packet a pair of consecutive rings was taken as a sample. Thuswe had observations of close gap on 216 Rings, 72 for each period, 54 for each Mandrel, 72 foreach packet consisting of 24 for each its three positions.

The averages of these are respectively shown in Figure F8.13 with appropriate ControlLimits. It is seen that:

25.19

25.24

25.29 USL

LSL


Blade No. 1 2 3 4 5 6 7 8 9 10 11 X

Y

RUN CHART 93


● The three periods conform to the control limits, indicating steadiness of process overtime. The three sigma control limits are approximately 0.015 Units apart. Since theseare limits for averages of 72, the variation in individual observations are estimated as0.015 (√72 ) = 0.1275 against Tolerance of 0.2, indicating a fairly high degree of processcapability, an index of over 1.5. Such a real situation, is often responsible forcomplacency or de-motivation, thus forestalling any urge to even think for improvementleave aside any serious attempt in that direction.

* Overall average of Rings produced from Mandrel �1� possess less Gap than thoseproduced from Mandrels 2, 3 and 4.

* The averages of positions of a packet show an increasing trend.* The averages of packets of a Mandrel also show an increasing trend.* The lowest average of about 0.415 corresponding to Packet A, position I almost assures

that the gap of individual rings from among these shall fall short of specified minimumof 0.40. This could have been avoided with process mean of 0.50 instead of 0.457.

Process : Oval O.D. Turning Machine: M Operator: O Facilitator: FModel: Ml Size: STD Parameter : Closed Gap Specification : 0.50 ± 0.10

Figure F8.13 Oval O.D. Turning.

It is thus clearly seen that actions on:* Proper maintenance of Mandrels (size and profile)* Proper setting for the Gap size and* Proper setting of Cutter Angle.together shall reduce the Biases arising from differences among Mandrels and the trends

between packets of a Mandrel as also among positions of a packet.Thus the Quality gets improved both in terms of reduced variation and closeness to

Target. Need, it be repeated that hitherto it was not considered necessary, since by and largethe Gap rarely violated the specified tolerances and the knowledge about the contribution tovariation from various sources was lacking. Graphical presentation of suitably collected datareveals both the potential for improvement and the possible approach to achieve the same.

0.42

0.43

0.44

0.45

0.46

0.47

0.48

0.49

0.50

0.51

1 1 1 112 2 2 22Periods

APeriods

CBPackets

Mandrels3 3 3 33 4

X

Process

Desired

Average

Average



Lastly a similar approach to monitor the �OD� of Piston in the same manufacturing unitresulted in Improvement in three phases. See Figure F8.14.


Phase: I showed large fluctuations around the mean which itself has been frequentlyshifting over time. 'The operators were educated to aim at the target instead of anywherewithin the tolerance limits. They were also advised not to reset the process in panic, thoughin good faith, but to wait for the signals from the limits, the trends and the associatedunnatural patterns.

Phase: II shows improvement over Phase I, yet there are runs of large number ofconsecutive points above and below the specified mean and several points on the border line,though no violations beyond the tolerances. The lessons given at the end of Phase I wererepeated and the desired cultural changes re-emphasised.

Phase: III shows excellent control, eliminating need for one hundred percent inspection.

ConclusionIt may be concluded from the above examples that a systematic approach to assess the

total variation and its components from various likely sources paves the way for rationalthinking of cost or effort to reduce variation from various alternative sources and alternative

–8–6

–4

–4

–4

–2

–2

–2

2

2

0

0

0

2

4

4

46

25

25

25

20

20

20

15

15

15

10

10

10

5

5

5

1

1

1

X

X

X

Lowertolerance

Lower tolerance

Lower tolerance

Uppertolerance

Upper tolerance

Upper tolerance

Y

Y

Y

Pha

se I

Pha

se II

Pha

se II

I

Target

Target

Target

RUN CHART 95


strategies versus gains in productivity of conforming products. By and large the extra costs, ifany, are negligible in comparison to the returns.

(d) Heavy electrical�Rotor Disc

The shape of the rotor disc looks like the Figure F8.15. Only 4 to 6 units weremanufactured in a year. This process was deemed non-repetitive and therefore considered notamenable to statistical process control. The job consists of drilling 48 holes. The job is loadedon the machine and appropriate settings made. Thereafter the same drill runs through all thesegments to make a hole. This process is repeated 48 times to complete the job. To astatistician this constitutes enough repetition to permit use of techniques of statistical processcontrol.

The problem is encountered in non-conformance of all of its parameters, such as diameter,pitch and concentricity of the holes. The data on all these parameters of all the holes (48 × 4= 192) were asked for in the sequence of their operation, to examine the pattern of variationamong (inter) and within (intra) segments. The management was serious about the problemand readily agreed to make available the data for the next job.

Figure F8.15 The shape of a Rotor Disc.

A lot many constraints and limitations did not permit data collection as envisaged andagreed upon. However data on distance �d� of 48 holes from the periphery in production



sequence was made available for the top segment, the cross section of which is shown inFigure F8.16.

Figure F8.16 Cross section of top segment of a RotorDisc showing location of 48 holes symmetrically placed.

Assuming that the periphery provides the correct reference with respect to the centre ofthe job, the distance �d� provides an index of the distances of its centres from the centre of thejob. With a pious hope that the plot of the data in the sequence of production might providesome vital clue to the pattern of variation, the data were plotted in a chart. The emergingdiagram is shown in Figure F8.17.

Figure F8.17 Run chart of distance ‘d’ of 48 holes of top segment of a rotor disc.

Y

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47Serial number of hole

43.5

44

44.5

45

45.5

46

46.5

47

47.5

48

48.5

Dis

tanc

e in

cod

ed u

nits

X

481

2

25

d

RUN CHART 97


It is seen that:* The difference between the largest value 47.9 and the smallest value 45.1 is 2.8 -

approximately 3.* If we try to draw a central line to the data, we find that the distance d has been

decreasing in the first half and increasing in the second half.* If we draw a pair of parallel lines, one on each side of the central lines forming a V

shape, wide enough to just include all or most of the points, we find that the distancebetween the parallel lines is about 0.75. This is approximately one fourth of theobserved maximum dispersion of 2.8 or about 3.

This leads to the conclusion that if one can trace the cause for this systematic trend andmake correction for the same, the total spread or dispersion can be reduced to its one fourthor its variance to just one sixteenth.

The shop was visited to have a first hand knowledge of the process, to look for the causeof the systematic trends - decreasing followed by increasing. The process consisted of placinga standard template on the top of the job and setting the drill to make hole through all thesegments, as positioned in the template.

It revealed that the template itself was not being placed appropriately, as per prescribedprocedure. The placement was not properly centred. It resembled the situation shown inFigure F8.18. This explains beyond any shadow of doubt, the root cause of the typical patternobserved. It needs to be recognised that no other tool, how so ever complex, can provide clueto this kind of lapse. Run chart is one of the simplest, most inexpensive (all one needs is apaper and a pencil), and yet most powerful tool to expose the unwanted trends that provideclue to the deficiency in the process and hence paves the way to the solution.

Figure F8.18 Comparison of desired versus actual fitment of template for drilling of holes.

Not only that, it was further seen that the template was designed to be located and fittedon the job with the help of two sets of nuts and bolts, placed diametrically opposite. Thepractice had been short cut and diluted to fastening only at one position on one side only, ingood faith. The good intentions failed to recognise that in the game or the race of qualityshort cuts are fatal.

Once the correct regime was followed, the spread further reduced to one third. Thus,overall the spread could reduce to 1/12th or the variance by 1/144th. The tolerance was aboutone sixth of the observed spread providing an index of 0.16 of process performance. This could

The template

The job and the template are not concentric.These are designed to be concentric.

The job



be enhanced to process capability index of about 2.0 making six sigma status a reality. Thisdid not involve any extra investment.

Likewise there are several cases where the assessment of capability in a judicious mannerand employing statistical process control compatible with its capability, in troubled areas haveresulted in elimination of nonconformities (reduction from about 60 percent to zero percent)and one hundred percent inspection.

The axiom that quality demands voluntary dedicated commitment to understand the rightway to do things and do exactly the same, to get the best possible from the available resources,gets reaffirmed. One needs to be passionate to give the best by habit. Habits die hard. Anyimprovements primarily coming from the new and supposedly better equipment is definitelyshort lived since we do not have the culture to sustain the way it ought to be run. We end upby adding to the costs without commensurate returns and increasing avoidable losses fromlack of quality many folds.

Perhaps, because of this precisely, it is said that quality is free. Passion for quality isspontaneous, it cannot be purchased or procured against money. In fact, Quality is cheaperthan free since no other investment gives such high returns. It ought to be remembered allthe time that doing things right does not cost even a paisa more while not doing right createsonly losses, sometimes substantial and even fatal. We are doing things anyway, all the time!Either right or wrong. Wastage of any resource is sin. The most precious and perishableresource is time. Therefore, no idling and no wrong doings. Of course the latter is worse thanthe former. Comparison is futile because both these habits are fatal diseases.

Run-Charts or suitable control charts as a means to monitor one's processes and use ofhistograms as a supplementary means for review with appropriate periodicity are simple,effective, efficient and indispensable tools at our service to proceed from symptom to causeand then on to remedy to fulfil the urge to improve continually to fulfil the needs of thesociety. There is no soft option.

8.6.3 Conclusion

Run chart provides a running commentary on the process behaviour just as ECG reflects onefor the heart. There is no substitute for this simple yet effective tool for the study of processbehaviour. Of course, coupled with the histogram at appropriate interval, it providesadditional value-able clues for symptoms leading to search of remedial measures. Theirconfirmation and implementation results in improvement of quality, productivity andprosperity with same resources and infrastructure. In exceptional cases the pros and cons ofinvestment required, if any, should be weighed against the likely returns, for taking rightdecision.

It has been experienced that in many cases, the processes considered incapable werefound to be capable or highly capable. The exercise of their control provide necessaryconfidence among the human resource and the product. This often led to replacement of onehundred percent inspection by nominal confirmatory sample checks.

In many cases maintenance of run charts on process parameters of concern at appropriatestages and compatible intervals ensures products of desired characteristics that alone shallguaranty its fitness for use. Such examples are common in foundries, textiles, vanaspati andhydrogenated oils, sugar, cement, plating, chemicals and what not. One needs to be innovativein deriving more and more from the usage of this procedure. The skill shall mature withexperience. True documentation and adherence to optimal standard work practices arenecessary to achieve excellence.

A COMPOSITE COMPREHENSIVE CASE STUDY 99


9A COMPOSITE COMPREHENSIVE

CASE STUDY

9.1 PROBLEM

The problem, faced by a heavy electrical manufacturing plant, required identification offactors causing increase in failure rate of turbo generator bars during high voltage flash testfrom 0.3 to 4.8 percent and thus to eliminate the failures or at least to restore the status quo.The cost of each bar in late 70�s of the 20th century was about a lakh of rupees. Thenonconforming bar is a total loss. It cannot be repaired or salvaged. At best it may sell as ascrap at nominal value. The incidence of failure or nonconformance, in this respect in theworks of the collaborators, is almost zero. These facts prompted the company to treat thisproblem on priority. This decision is based on the PARETO principle of choosing the problemthat hurts most.

9.2 APPROACH

As a first step, a meeting of all technical experts from production and associated services likeDesign (product), Process technology (methodology), Maintenance, Testing was organized. Justby way of information the flow diagram is shown in Figure F9.1.

The brain storming session led to listing of the following 16 likely factors or causes orsources effecting the performance of turbo generator bars:

(i) Season (Month)�Environment (Temperature)(ii) Shift of taping of insulation (Degree of supervision influenced by the team of

executives and supervisors (technical personnel present)(iii to viii) Taping parameters�Dimensions�Height, Width and Periphery. Each Before

and After curing�that is once after taping and again after curing with theidentity maintained.

(ix) Time lag between taping and curing stages of the process(x) Shift of curing

(xi) Different types of Moulds(xii) Calibration of moulds�Pre and post calibration effect

(xiii) Delta tan delta (∆ tan δ)(xiv) Time to fail (from start of testing or switching on to the failure of the failed

bar)



(xv) Location of failure(xvi) Age of the insulation tape (the interval between the date of manufacture of

the tape and its usage namely taping)These are presented in the form of a CAUSE AND EFFECT DIAGRAM in Figure F9.2.The process data had been recorded satisfactorily since inception, at respective stages. It

was easy to transfer the same to one Master CHECK SHEET, shown in Table T9.1, developedfor this purpose, for the desired period.

Even though, it is advisable to first look into the past data, as thoroughly as possible, tomake a head start, yet as a short cut the use of techniques of Design of Experiments andMultiple Regression were contemplated. Inspite of the keenly interested involvement by themanagement, these two approaches could not be executed, for one reason or the other, such asconstraints on existing facilities for testing and control, production rate and the productionprocess. After loosing precious time, the team fell back on the first option, namely to examinethe available past data.

9.3 DATA PLANNING

To make optimal use of the past data, it is desirable to plan its quantum and period.To study the effect of season or month [cause (i)]: A minimum of two years data, month

wise, are necessary. Latest two year data were compiled.For causes (ii), (ix) to (xii), (xv) and (xvi), which are of attribute kind, it was planned to

use data for the latest one year only.For the causes (iii) to (viii), which are of measurable or variable type, it was planned to

collect data for three months only; the best, the average and the worst months of the latest

Sanding finishing & varnishing

Final curing in mould

Curing in machine

Brazing of lugs

Transposition

Solid copper conductor

Hollow copper conductor, glass tape, varnish Hollow conductor taping

Conductor cleaning, cutting and end shaping

Placing 28 solid and 14 hollow conductors in 2 layers

Pressing slot portion & cotton tape removal

Over hang formation

Removal of fallow & making the surface even

Assembly teflonand cotton taping

Periphery measurement

Count of number of layers of tape

Assessment of tan & H.V. testD d

Taping in air-conditioned room

Measurement of height & width

Packing & despatch to assembly

Assembly

Figure F9.1 Flow diagram of the process of manufacture of TG Bars.



Figure F9.2 CAUSE AND EFFECT DIAGRAM for failure ofturbo generator bar during High Voltage flash test.

Table T9.1 Master CHECK SHEET.

Supervisor

Supervisor

Supervisor

Year:Month:Date:

Width:Height:

Shift of taping:Date of manufacture of insulation tape:Age of tape (days):Before taping-periphery:After taping-periphery:

Shift of curing:Interval between taping and curing [No. of shifts]Mould type used: single/twin, lower/upperMould calibration date:

Testing: failure location:Time to fail (sec.):Voltage; tan :D d

Bar number:

Company: Department:

Age

Insulationtape

Shift

ShiftProcess

Interval between taping& curing (No. of shifts)

CA

AWidth

AfterHeight

Periphery

Periphery

Taping

BeforeNo. of layers

Curing

PressureTemperature

Handling

B

B

MaterialLocation

(Proneness)Season(Month)

Left

Corner

RightCentre

After

Calibration

Before

Lower

Lower Upper

Upper

Type

Single

Twin

Mould

Not considered–past data not available

Failure ofturbogenerator

bar

Duration



one year, as examined for the attribute parameters mentioned above. For each of thesemonths, data for 25 bars only selected at random from the month�s production werecontemplated. The taped stuff is soft. Therefore, the measurement of height and width at thisstage is subject to large error. These dimensions, therefore, are not recorded. Themeasurement of periphery is less erroneous and is considered good enough index of theamount of taping. After curing, however, all the three dimensions are recorded. All thesedimensions play an important role during assembly, too.

For the remaining two causes xiii and xiv, it was feasible to collect data on failed barsonly. Data on Time to fail with corresponding information on Delta tan delta and Age of thetape used were documented for a representative sample only.

The data maintained were considered satisfactory and therefore no fresh data wereplanned.

9.4 ANALYSIS AND CONCLUSIONS

Let us examine, the effect of these one by one.

Cause i, season or month:

Figure F9.3 RUN CHART of failure rate of TG Bars for twoconsecutive years, month wise, to assess the effect of season.

Monthly RUN CHART of percent non conformance, in the form of line graph, for the twolatest consecutive years, is shown in Figure F9.3. It is seen that in one year the period Aprilto August is worse than the next seven months September to March. For the other year thepicture is just the opposite, the first five months are relatively better and the next sevenmonths worse. In case of any worthwhile dominance of the effect of the season or the month,the two annual graphs should be either congruent or at least parallel. In fact, if any thing, thereverse trend between the first five months and the next seven months, for the two years,contravenes the conjecture of any systematic effect of the months, beyond any shadow ofdoubt.

Apr. May. Jun. Jul. Aug. Sep. Oct. Nov. Dec. Jan. Feb. Mar.

x0

2

4

6

8y

% fa

ilure

Month

Year-Y

Year-Y + 1



This finding, was not relished by the technical personnel, in the organization. They hadliterature from their collaborators stating that seasonal environments, temperature inparticular does contribute to the incidence of failure during high voltage flash test. It tooksome time and effort to explain to them that the real implication of the finding is that, thecontribution to the incidence of failure from some other factor(s) is relatively so high that thecontribution of seasonal effect, if any, is relatively too small to be perceptible. In other wordsthe effect of season, if any, is so dormant that it is camouflaged to such an extent by othermore dominant factor(s), that it has become imperceptible. If the effect of dominant factor(s)is(are) taken care of by elimination or adequate reduction, then its effect might show up. Thepicture at the collaborators could well be different. The two situations are not exclusive toeach other.

Cause ii, Shift of taping of insulation:

The taping of insulation is done in two shifts designated as A and B only, to the extent ofproduction needs. The data on production and nonconforming number (failures), during thetwo shifts are presented in Table T9.2. This process is termed STRATIFICATION orclassification by shifts.

Table T9.2 Nonconforming T G Bars STRATIFIED by shifts of taping.

It is seen that a total of 959 bars were produced during the year. 32, of these were foundnonconforming. These failed during high voltage flash test. Thus, on an average, thenonconformance was 3.34 percent. Out of 959 bars, 735 had been processed during the shift Aand the remaining 224 in shift B. The respective nonconforming units were 26 and 6 or 3.54and 2.68 percent only. Now, let us take a bowl containing 1000 beads, 967 green and 33(3.3%) red. Green represent good bars and the red the nonconforming bars. If, we pick up asample of 735 beads at random, from the bowl, it is likely to contain 24 red beads in the longrun on an average, if we repeat the process a large number of times. However, in any singletrial, the number will vary marginally, this way or that way, due to what we may call asampling error.

Shift Number

Produced Non-conforming

Non-conforming

% *

* If d exeeds 2 but does not exeed 3, it is considered to have some mild influence.If d exeeds 3, it is considered likely to have worthwhile influence.*

IdI

5

0.20

0.66

0.00

S

6

0.66

1.20

0.45

7

0.30

0.55

0.00

4

3.54

2.68

3.34

3

26

6

32

2

735

224

959

1

A

B

Overall

P = 3.34, Q = 100-P = 96.66, PQ = 322.84, s = (PQ/n)IdI = deviation s = standard deviation

d in unitsof s



It will be nice to play this game or experiment to become familiar with or become awareof or gain first hand knowledge of the likely magnitude of sampling errors, which are verymuch a part of our day to day life. Soon, we shall realize, that the above deviations ordifferences for the situation on hand, namely samples of sizes 735 and 224 with respectivenonconforming units of 26 and 6, are reasonably small and can, therefore jolly well occur asa normal phenomena and therefore can be given benefit of doubt as arising from samplingerrors.

In fact, if we make appropriate calculations and refer statistical tables or ready reckonerfor the purpose, the deviations of the magnitude, observed above, are likely to be exceeded onabout 80 and 60 percent of the occasions respectively. For 95 percent confidence level, evendifferences up to twice this magnitude could be exempted. Therefore, there is no justificationto look for the reasons for this difference between the shifts, as long its magnitude does notexceed the prevailing status. Any efforts to look for difference of this magnitude, are morelikely to prove futile in our endeavour to find a solution to the problem on hand.. This way wecan optimally use our human resource more effectively. We avoid spending our energy onfutile exercises and conserve it for use on better occasions, when our mind is fresh to respondto the demands of the problem. On the contrary the matter is worth consideration if thedifference exeeds the natural limit.

The executives did not consider this finding to their taste. Since, this process runs in twoshifts namely, A and B only, when executives are present in varying number. They relentedand suggested its examination over three shifts. The curing process extends to three shifts. Itsexamination is a part of this study. It is discussed under cause x, the shift of curing,subsequently.

Causes iii to viii�Taping Parameters�Periphery before curing, Height, Width andPeriphery after curing:

As explained earlier, the latest year data were used to identify the three periods�thebest, the average and the worst months. From the produce of each of these three months, arandom sample of 25 bars was identified. The data on; the periphery, both before and aftercuring; and height & width only after curing was culled out for all the three periods. All thedata thus obtained were summarized in twelve HISTOGRAMS. These are shown inFigure F9.4.

It is seen that:* the histograms for height and width, after curing, for all the three months worst,

average and the best are quite akin to each other.* the histograms of the periphery after curing exhibit reduced spread in relation to those

for the periphery before curing for all the three periods. This is expected to be so. Thestuff is soft before curing and hard after curing. The measuring error is likely to behigher in the former case.

* the histograms of the periphery before curing show trend for successive reduced spreador variation over the three periods.

* the histogram of periphery after curing for the third period has distinct single peak incontrast to double peak for those for the previous two periods.

It may thus be concluded that if at all, the periphery of taping may have some influence onthe quality of T G Bars with regard to its fitness to survive the high voltage flash test.



Figure F9.4 HISTOGRAMS of variable parameters of insulation of Periphery beforecuring and Periphery, Height & Width after curing for the 3 months, worst, average and best.

Time lag inshift

Number


Non-conforming

%IdI

5

0.69

0.92

0.00

s

6

0.77

0.88

0.45

7

0.90

1.05

0.00

4

4.03

2.42

3.34

3

22

10

32

2

546

413

959

1

< 2

= or > 2

Overall

d in unitsof s

Table T9.3 Nonconforming T G Bars STRATIFIED byTime lag between the processes of taping and curing.

Cause ix�the time lag between the processes of taping and curing

The data are suitably presented in Table T9.3.This is STRATIFICATION by the duration of waiting. The dust is considered to be enemy

number one of the electrical and electronic goods in process. The longer the waiting, the moreare the chances for gathering the dust.

0

0

0

20

20

20

40

40

40

60

60

60

Worst

Average

Best

Fre

quen

cy p

erce

nt

y

Before curing After curing Month

Periphery Periphery Height Width

176

180

184

188

223

225

227 86 88 90 26 28 30

Dimensions in mm

x



Using the same logic as discussed for cause ii, the difference of this magnitude betweenthe two strata can be given benefit of doubt as having arisen from the sampling errors. Thedeviations of this magnitude could have been exceeded with about 40 and 30 percent chances.It is therefore of no consequence. This implies that longer waiting does not contribute toexcess nonconforming units, nor does it imply that it contributes to its reduction. Longerwaiting, needs to be discouraged for its adverse effect on goods in process and theconsequential productivity.

However, numerically, the incidence of nonconformance is higher for lesser waiting. Thisassures us that if due to some extraneous factors, the waiting time gets longer, one need nothave tension for unfounded fear of its likely adverse effect on failure during high voltage flashtest. The concern would be different, if longer waiting were seen to be harmful. It alsosufficiently, though indirectly, indicates that the storage environments in the air conditionedprocessing area are absolutely satisfactory.

Cause x�the shift of curing

The curing of the T G Bars is done in three shifts designated as A, B and C only, to theextent of production needs. The data on production and non-conforming number, during thethree shifts are presented in Table T9.4. This process is termed STRATIFICATION orclassification by shifts.

Interpreting the data in the same manner as for cause ii, it is seen in this case that,unlike as observed for the shift of taping there is an evidence, to the extent of about 98 %level of confidence, of deterioration during shift C. The deviations for the shifts A and B couldexceed the observed magnitude by about 10 and 85 percent due to chance. The performanceduring shift C is almost twice as bad as the average of the three shifts. The hunch of theexecutives that when the cat is away the mice play or the diluted supervision during the shiftC may have an adverse effect on the failure during high voltage flash test, is not completelyunfounded. There might be some truth in it.

Table T9.4 Nonconforming T G Bars STRATIFIED by shifts of curing.

Cause xi�the types of moulds

The nonconforming of T G Bars STRATIFIED by the types of moulds in use are shown inTable T9.5.

Using the same logic as in the interpretation of the impact of cause ii, it is seen that thereis evidence, to the extent of 2% level of significance or 98% confidence, that single upper

ShiftNumber


Non-conforming

%IdI

5

1.38

2.78

0.20

0.00

s

6

0.84

1.15

1.13

0.45

7

1.64

2.42

0.18

0.00

4

1.96

6.12

3.14

3.34

3

9

15

8

32

2

459

245

255

959

1

A

C

B

Overall

d in unitsof s



mould might be contributing to higher rate of failures marginally. The deviations showing upfor the rest of the moulds could have exceeded the observed values with the chances of about20, 30 and 60 percent respectively. This Indicates that single upper mould might be a fit casefor calibration.

Cause xii�impact of calibration of moulds

To study the presumable salutary impact of calibration on the failures observed duringhigh voltage flash test, the calibration intervals were looked into. It was consideredappropriate to STRATIFY the failures of T G Bars by their incidence during immediatepreceding three months of the calibration date and succeeding three months. The data areaccordingly presented in Table T9.6.

Using the same logic, as for, cause ii, it is inferred that the act of calibration did not doany good to the process. The performance was about the same in both the situations, namely

ShiftNumber


Non-conforming

%IdI

5

1.18

1.82

0.69

2.54

0.00

s

6

0.93

1.56

1.46

1.03

0.45

7

1.27

1.17

0.47

2.47

0.00

4

2.16

1.52

2.65

5.88

3.34

3

8

2

4

18

32

2

370

132

151

306

959

1

Single lower

Twin lower

Twin upper

Note :Mould single upper significant at 2% lavel.It was checked since its cost is economical.

Single upper

Overall

d in unitsof s

Table T9.5 Nonconforming T G Bars STRATIFIED by mould type.

Table T9.6 Nonconforming T G Bars STRATIFIED by pre and post calibration of the moulds.

Calibrationstatus

Number


Non-conforming

%IdI

5

0.51

0.41

0.00

s

6

1.45

1.28

0.96

7

0.35

0.32

0.00

4

4.21

5.13

4.72

3

9

14

23

2

214

273

487

1

Pre

Post

P = 4.72, Q = 100-P = 95.28, PQ = 449.722, (PQ) = 21.207, s = (PQ/n)

Overall

d in unitsof s



before and after calibration. This implies that the decision to calibrate the mould was taken inpanic, when one was not able to pin point a specific cause for increased incidence of failuresduring flash test. The action is taken on the basis of results during the test and this may notnecessarily be linked to the current status in the curing area.

Causes xiii and xiv�The impact of the value of ∆ tan δ of the bar on the time elapsed tillfailure

The data were available on the ∆ tan δ value of the bar and its corresponding time to failin seconds since switching on the current, for the failed bars only. It was decided to look intothe relationship between the two, if any. For this purpose a SCATTER DIAGRAM wasattempted, with the former on the x-axis and the latter on the y-axis. The outcome is shownin Figure F9.5.

Figure F9.5 SCATTER DIAGRAM of time to fail versus its ∆ tan δ value of the failed bars only.

It is un-fair, to look for the relationship for the truncated data, namely for the failed orthe nonconforming bars, only. The constraints of the available data did not permit the look atthe total picture. But with a view to salvage whatever information was possible, it wasthought pertinent to examine whether the parameter delta tan delta influences even the timeto fail.

The look as the Figure F9.5, shows that, by and large, the bars which have taken bothless or more time to fail are also associated with both low and high values of delta tan delta.Thus in the light of above two considerations it is unfair to infer any association of ∆ tan δvalue of the bar, in its present range of variation, with its failure from the available data.

Cause xv�location of the failure

It was considered worthwhile to examine, whether any particular location(s) or segment(s)of the bar was(were) more prone to fail than the rest. If so, it might provide some clue(s) towork for the remedy. This art of looking at the concentration areas of the incidence of interestis called geographical STRATIFICATION. The outcome of this exercise is shown inFigure F9.6

It is seen that the portion along its flat length, which is taped mechanically, is free of anynonconformity. Out of 32 nonconforming units, nine are located near the left bend and the

0 5 10 15 20 25x

20

40

80

y

Tim

e of

fail

(in s

ec.)

D dtan

(Failed bars only)



remaining twenty three are concentrated near the right bend. It may be mentioned that thetaping near the bends is manual. Thus we may infer that there is a possibility that manualtaping near the bends is prone to be more weak, than mechanical taping in the flat portion,resulting in the failure of the bars and more so near the right bend than the left bend.

Figure F9.6 Geographical STRATIFICATION, showing concentration of location for theincidence of failures, along the profile of the bar, rendering it as nonconforming.

Cause xvi�Age of insulation tape

Lastly, the failures STRATIFIED by the age the insulation tape split into nine classes isshown in Table T9.7.

200 500 2940 100

15 64

239

16

200500

Nil

Number ofbars failed

Slot portionin mm

Over hangexciter side

Over hangturbine side

Table T9.7 Nonconforming T G Bars STRATIFIED by the age of the insulation tape used.

Number


Non-conforming

%Age(in days) IdI d in units

of ss

1

<50

110-129

50-59

130-149

70-89

170-189

90-109

170-189= or > 190

Overall

2

12

126

113

97

138

141

130

16735

959

3

0

6

0

5

0

7

2

102

32

4

0

4.76

0

5.15

0

4.96

1.54

5.995.71

3.34

5

3.34

1.42

3.34

1.81

3.34

1.62

1.80

2.652.37

0.00

6

5.19

1.60

1.69

1.82

1.53

1.51

1.58

1.393.04

0.45

7

0.64

0.89

1.98

0.99

2.18

1.07

1.14

1.910.78

0.00

393

566

2

30

0.51

5.30

2.83

1.96

0.91

0.76

3.11

2.58



It is seen that the failure rate is zero or nil as long as the insulation tape is less thanninety days old. It is 1.54 percent (2 out of 130) when the age is 90 to 109 days, both endsinclusive. It needs to be understood that if the lot contains zero nonconforming unit or non-conformity, the sample too shall contain none of these. On the contrary, if the sample containszero nonconforming unit or nonconformity, it is not necessary that the lot too shall containzero nonconforming unit or nonconformity. If we club the first four classes, we have 2 failuresout of 393 giving a percentage of 0.51. With 0.51 as the average percent nonconforming, thesamples of sizes 12,113, 138 or even 262 (= 12 + 113 + 138) may contain zero nonconformingunit or nonconformity. Giving this phenomena a chance, we club these four classes as alike.Like wise the remaining five classes may be considered alike with percent nonconformingunits as 5.30 (30 out of 566). Dividing the entire data into two groups namely, age less than110 days and 110 days and above ensures that the expected number of failures based on theoverall performance is not less than five�a criteria for making rigorous valid tests forsignificance. The incidence of 0.51 percent based on the sample or production of 393 bars doesnot contradict, the past stable percent of non conforming units of 0.3 percent. The deviation ofthis magnitude can safely be attributed to the sampling error.

To sum up, the age of the tape at the time of taping, is single most important factorinfluencing the failures during high voltage flash test. The other four mildly suspectedcontributory factors or causes; namely periphery of insulation tape after taping, shift ofcuring, single upper mould and the location near the bends; vanish into thin air, if only thetape used is fresh. Use of fresh insulation tapes, within three months of its manufacture,makes the process robust to take care of unintended imperfections arising from the fourfactors mentioned hitherto, inclusive of the skill of the workers doing manual taping in thecurved zone of the T G Bars. To be on the safe side one may decide to use the tape withinthree months of its manufacture. The useful life at the collaborators end is one year.

9.6 THE SHELF LIFE OF THE INSULATION TAPE

The shelf life is a crucial factor. The tapes are required to be stored in air conditioned roomand that too within specific temperature band. The storing area of the tapes was audited. Itwas found to be satisfactory. No flaw whatsoever was perceptible. The matter was probedfurther. The tapes are requisitioned from collaborators abroad. They intimate, the dispatch byair, as and when the supplies are made. The consignment is received at the airport by theteam equipped with trucks full of bricks of ice, to be transported to the plant site about 200kilometers away. There were occasions when the plane carrying the consignment is delayedand the availability of ice is not adequate. Even if sufficient ice were on hand, the vagaries ofthe environment may not permit sustaining the desired temperature in stipulated range ofhumidity. The problem thus reduces to improving conditions during transit of the insulatedtapes to enhance its shelf life to one year or use it within three months or strike acompromise between these two extremes.

9.7 RECOMMENDATIONS AND CONFIRMATORY TRIALS

It was recommended to:● use fresh tapes with shelf life less than 110 days.● check single upper mould.



● maintain desired periphery during taping operation through adjustment or control ofnumber of layers of the insulation tape.

● educate the workers on the why of conforming to product and process parameters orthe tangible & intangible loss that their nonconformance can cause and the processknowledge of how to assure desired conformances. The workers need to be adequatelyfacilitated and empowered.

As a long term strategy, consider measures to prolong the shelf life of the insulationtapes.

Immediately the confirmatory trials were started with the insulation tapes of shelf life upto 109 days. The older tapes were used for other purposes where requirements were not sostringent. During four months of this regime, 324 T G Bars had been produced. A total of fourbars had failed the test including the two that got damaged in an accidental fall. Excludingthese two, the failure rate works out to two out of 322 or a percent of 0.6. This is consideredcommensurate with the past best stable performance of 0.3 percent against the deterioratedlevel of 4.8 percent. Thus the old good status had been restored. The potential to achieve zerononconforming is not a distant dream with another couple of iterations.

9.8 CONCLUSION

Thus the exploitation of the simple tools depends upon the ingenuity, skill and the technicalknowledge of the user or the using team about the problem area. The advantages areboundless. It does require:

● element of preplanning of requisite relevant data with necessary traceability,● collection of adequate (neither more than what is necessary for sufficient evidence nor

less than that will keep the suspense on the conjectures or the hypothesis alive) dataeconomically,

● data to be true.This should be followed by prompt valid analysis or summary leading to recommendations

that need to be confirmed through trials before implementation on routine bases. The gainsaccrued need to be sustained before attempting another breakthrough.

All these efforts and allied inputs are investments that pay rich dividends. The questionis not whether we can afford these investments but whether we can do without these? Thegains are made possible through commitment to quality or the needs of the customer. This isnot possible through money alone. Of course, the invested seed money comes back many timesover. The way to survive in this competitive world is putting one's heart and soul into one'smission qualitatively.



10ORGANISING FOR QUALITY

9.1 NEED FOR QUALITY

Quality is something with which we are all concerned. Quality, of product, services andwhatever we do, influences us all, in one way or the other, at all times and every where, fromour health through activities to prosperity and even happiness.

Lack of quality impacts all of us. Tardy growth rate, rapidly growing unemployment andgalloping inflation are some of its consequences. During the last three or four decades theexchange rate of Yen, the Japanese currency, has strengthened several folds in relation to USDollar, while that of India has weakened several folds. The Yen and Rupee ratio has taken asteep dip.

Sloppy quality costs more through losses it inflicts. We suffer impact of poorcommunication among host of other essentials. The inefficiency of services is invariablyattributed to poor quality of products or its improper use�both provide symptoms of lack ofquality culture. Non-availability of funds is in turn offered as a convenient excuse.

Quality of product and services is indispensable for our survival.

10.2 REWARDS OF QUALITY

Against the background of losses caused by the lack of quality, the following have beenexperienced in Indian industries.

One cable manufacturing company estimated that in terms of economic viability, onepercent reduction in nonconforming units of the product was equivalent to augmentingproduction by ten percent, which otherwise would require more investment in space,equipment and human resource. The alternative course of reduction in nonconformancethrough prevention of non-conformities offers a viable economic proposition. This motivatedthe human resource to bring nonconformance to one twentieth of the prevailing rate and thusensure huge recurring savings at no extra cost.

Nonconformance of batches of disinfectant requiring reprocessing was reduced from thirtyfive to one percent in an insecticide production unit.

Rework of varying degree on all characteristics of a component was completely eliminatedin an engineering organization facilitating assembly and enhancing product quality.

Nonconformities of over thirty percent during manufacture of blades of turbine in a plantmanufacturing heavy electrical equipment was reduced to under one percent, in a span of

ORGANISING FOR QUALITY 113


about four weeks. This enhanced absolute production three folds and conforming productionfour folds.

Some state transport undertakings are providing better services at lower cost than someothers.

Scrap generated during production in an electronic factory was reduced to one tenth ofthe prevailing rate.

Insulation property of ceramic insulators was improved to acceptable norms.Admixture of glass particles in sealed bottles of a baby tonic were completely eliminated,

assuring its safety for the humans.The efforts needed for above achievements required little or no investment, yet yielded

abundant returns, that could build up resources for further development at a faster rate.Contrary to common belief that improving quality requires more costly inputs and

reduced production, it instead begets enormous tangible and intangible rewards.All these confirm that the national potential to achieve quality level of zero non

conformance or parts per million or six sigma status exists. That too, at no extra cost or ateconomically viable cost. Can we afford to let go, this potential?

10.3 QUALITY IS THE BEST POLICY

From the foregoing we conclude that there is no substitute for quality. It pays in all situations,be it times of plenty or scarcity, inflationary or deflationary, protective, monopolistic orcompetitive market, since it always adds to profit by reducing manufacturing costs throughprevention of nonconformities and avoidance of waste of resources in all its forms. The poorquality of work culture generates losses through enhanced rework, scrap, handling, productionloss, customer dissatisfaction and associated intangible harm. The salvation from these illscan be achieved by doing every thing right first time. By preventing nonconformities weactualize economic control of the quality of manufactured products and services provided. ThusQuality, is no doubt, an ALL WEATHER FRIEND.

This leads us to the dictum, take care of quality and quantity will take care of itself,converse is not true. Hence as a policy, assign top priority to quality.

10.4 ESTIMATE OF POTENTIAL SAVINGS

An approximate idea of avoidable losses can be had quickly by the following methods ofassessments.

Estimate the ideal minimal cost of product, likely to be incurred, assuming zero lossarising from wastage of material, under utilization of equipments and facilities includinghuman resource. Compare this with the actual cost being incurred. The gap is the loss or thepotential diamond or even platinum ready to be grabbed.

Estimate through simple work sampling study the ratio of activity time spent on doingfresh job to time spent on non fresh jobs in making the same worthy of acceptance. Divide thetotal cost of the infrastructure in the same ratio. The latter provides the tentative estimatebeing sought.

Compare consumption of each input like material, machine hour, man hour, capital etcper unit of conforming output over convenient periods for the recent past. Take the least of



each of these as the achievable and work out the least possible, with these norms. Thedifference between this and the current cost gives us the estimate we are looking for.

Ultimately, one needs to organize data base that shall permit estimation of variouselements of quality cost as per IS : 10708�1985 for review to plan and execute strategies forcost reduction. Fifty percent of the prevailing avoidable losses may be taken as an estimate ofthe immediate target of potential savings. It is invariably several folds of the profits beingcurrently realized.

10.5 FUNCTIONS TO BE ADDRESSED

Attainment of quality cannot be left to chance. The cycle of improvement and maintenance ofquality needs to be sustained through well planned and executed systematic activities whichneed to be updated through periodic review at appropriate intervals, depending on thesituation on hand.

The activities may be divided among three functional groups�PREVENTION,ASSURANCE AND IMPROVEMENT.

PREVENTION through quality control measures:

Associated activities consist of

Assessing the needs of the customer and deploying these functionally.

Assessment of process capability and its optimal exploitation. Reference may be made toIS : 10645 and IS : 397 parts 0, 1, 2, 3 and 4.

Choosing the right suppliers.Experimentation to develop fool proof product and process designs with minimal

necessary effort and cost that assure valid reproducible results. Reference may be made to IS:10427 parts 1, 2.

Planned collection of data and their appropriate analysis. Reference may be made to IS :7200 parts 1, 2 and 3.

Training in statistical control methods for improvement and maintenance of quality ofentire human resource in all segments and at all levels or as often said, both horizontally andvertically. Specially tailored syllabus is structured for each professional group. For example,the operators need to be educated particularly on product and process knowledge includingremedial measures to cope with deviations that may arise, even though occasionally.

Inspection as a means for prevention, ASSURANCE and audit:

Inspection is not quality control. It provides important means for the same when itsresults are used to locate the sources of nonconformities to avoid their recurrence. It includesactivities like:

Assessment, reduction, standardization and control of non-sampling inspection errors.

Selection of standard inspection plans to economically contain consumer's and producer'srisks.

Inspection of incoming materials, in process and finished goods.

Inspection, during the process by the task performer himself, plays the role of preventionof the nonconformity and quality assurance. Such inspection takes three forms. These are firstoff inspection, patrol inspection and last off inspection.



● The first off inspection consists of inspecting the first unit or a small sample ofconsecutive units produced, after satisfactory completion of all preparatory work. Thisis done to confirm or ensure that the process starts right. Well begun is half done. Incase of the unlikely event, that the first piece is not satisfactory, the processparameters are looked into to readjust them to the right levels before formal activationof the process.

● Having started right, the next step is to keep right. This is done through patrolinspection consisting of inspection at suitable regular intervals. The process is notdisturbed as long as its produce is satisfactory. The recommended interval is half of theexperienced stable running period of the process. As soon as the contrary event isencountered, the process is stopped. It is restarted only after satisfactory correctionshave been made. Also the produce since the previous inspection is segregated toprevent a nonconforming unit reaching the customer. All the nonconforming units sofound are acted upon in a manner to minimize the loss. This way, the patrol inspectionhelps us to continuously run the process rightly.

● Lastly, we inspect the last piece to ensure that the entire produce is satisfactory. Notonly that, the set up is dismantled and all the jigs and fixtures used during theproduction are checked for their fitness for use before crediting to the stores. Theseactivities constitute last off inspection. This avoids surprises often encountered whenthe production of this item is taken up next. Thus firefighting that hurt the systembelow the belt are avoided. These result in financial savings as tangible gains andpeace as intangible bonus.

The above activities together constitute process control in action. This way alone, doesinspection play the role of quality control. Or else it does only the post mortem. Theinspection of a nonconforming unit is synonym with postmortem of the dead body. Often theloss or the harm that a nonconforming unit can cause is many times more. For example, anonconforming unit of an automobile may cause an accident and many fatalities.

For salutary effect, the operator implying the entire human resource needs to be in stateof self control. He needs to know the why of the product and how of the process. He shouldhave means to know what his happening. He should have the knowledge and means to correctthe process, in case an unwanted deviation appears. He should be adequately facilitated withnecessary equipments and empowered.

* Maintenance and calibration of all inspection instruments, gauges, test equipments,tools, jigs fixtures etc is equally essential to accomplish the above functions as intended.

Assurance through Audit include activities such as:

Check inspection, Accuracy of inspectors, Analysis of customer complaints, survey of usersexperience, Assessing customer satisfaction and market quality determination.

Executive reports on quality with appropriate coverage, frequency and circulationThese activities ensure that the well thought out quality plans are adequate and in place.The planning of all activities associated with the above functions may be centralized and

execution decentralized.For result oriented efforts, it is necessary to synchronise activities of all concerned. These

should aim at achieving company goals. A company manual on quality serves a s a unifyingforce. It fulfills the following needs:

● Reference to policies and procedures with details of what, when, where, who, whom,why and how. This necessitates good advance thinking and assures accompanyingbenefits.



● Text for training.● An aid to continuity of operations despite employees turnover.● Basis for audit of current practices.● Precedent for future decisions.Lastly IMPROVEMENT is key to growth, the lack of which is fatal. All improvements are

made possible through project by project studies. The group of workers accomplish the taskwith support from inter descipline teams and leadership from top management which ensuresnecessary facilities and empowerment.

10.6 INPUTS FOR QUALITY

Quality has been defined as fitness for use. This lays emphasis on total satisfaction of the userof the product or services. It encompasses:

Appearance : Good appearance makes first sales.Functional : expected performance generates repeat sales,Reliability : implying economic availability, maintainability and effective life cycle cost,

sustains the growth of market share.All these aspects are translated into specifications for material, process and product

compatible with the required usage. Conformance to these standards in turn will assurefitness for use. It is an upward moving target. It is imperative that means to measure orquantify these are updated. These should be adequate, just good enough for the situation,neither too good that makes it more expensive and in-competitive in the market nor deficientthat defeats the very purpose.

The input-output cycle is described in Figure F10.1. All inputs excepting men includingwomen or human resource are inanimate. The wars have proved, repeatedly, if proof wereneeded, that it is the man behind the machine that matters. A well educated, trained, skilledand committed HUMAN RESOURCE of high integrity can assure appropriateness of all otherinputs. Therefore the entire human resource besides being in state of self control in respectivework area, ought to be adequately facilitated and empowered. Thus training and retrainingassumes vital significance. One of the key factors for the success of the Japanese in industrialrevolution has been massive education particularly in simple statistical methods in qualitycontrol on radio and T V network.

In India too use of mass media in education of human resource engaged in agriculturethrough programmes like krishi darshan has transformed the era of food scarcity to that ofplenty. So much so, that other countries wanted the details of its curriculum. India rightlygave first priority to food front for survival. Now, we can look forward to industrial revolutionfor prosperity.

Success in this venture should be easier, since relatively industrial human resource ismore literate. Even though massive efforts are involved on several fronts to reach the entireindustrial human resource both horizontally and vertically in all segments, the mission ofleading revolution to usher era of prosperity is reachable in foreseeable future.

10.7 TOTAL COMPANY WIDE ACTIVITY AND SELF QUALITY CONTROL

Activities need to be total and company wide that involve every one with mutualunambiguous communication and feed back based on strong factual data. Every individual



should understand his role, do it religiously and keep it updated. This habit cannot be forcedupon. It has to be voluntary, natural and self driven or motivated. This brings us to theconcept of self quality control. It implies that, each one should understand one's role andperform it accordingly right first time. Of course, the human resource needs to be suitablyfacilitated and empowered for the success of the company wide mission of quality leadership.

Our inconsistent attitudes spell disaster for quality. One when we are responsible toproduce goods or providing services and another when we are recipient of the same. Theselead only to complaints and in house firefighting. A case, in example is, that millions of usworking in transport service blame an equal number among us working in telecommunicationservices for its poor efficiency and vice-versa. Each group has many workable advices for theother. Remedy lies in each group organizing and practicing requisite measures inaccomplishing the goals in respective work areas for the satisfaction of us all vitallyconcerned. How can one receive good product or service unless some one provides it? Charitybegins at home, so does service start with self.

10.8 PRACTICE QUALITY FOR PERFECT QUALITY

It is well documented that there exist many situations in every plant which are not manifestbut are causing avoidable colossal waste of resources. What needs to be recognized however isthat there exists at least one corner in every plant, which is a symbol of excellence. Healthypractices of this corner need to be traced and emulated as a standard by all concerned, tosuccessfully combat ill effects of the prevailing bad habits, to be worthy of leadership.

10.9 IS/ISO : 9000 FAMILY OF STANDARDS

It is a noble document. Its advice is simple, understand the needs of the customer includingdetails of:

What does he want?How much does he want?Where does he want?How much worth does he thinks it is? andWhen does he want?and go all out, intelligently to delight him viably.The standard makes the use of statistical methods obligatory and mandatory.It is experienced that managements going in for certification, find opportunities and or

excuses to escape their use. On the contrary they should be looking for opportunities whatelse and where else can the simple statistical methods be used. The organization shouldimbibe statistical culture to own the quality culture. It should be using these tools because ithelps them to perform better and better. The continued effort can transform them to be aleader. Then they are worthy of the certificate. Instead there is craze to manage to get thecertificate as quickly as possible without necessarily developing the habit of improvements ormastering the culture expected of the organization by the standard.

Let the noble document intended as a guide be treated as such, with the reverence itdeserves in one�s own interest and the interest of the society at large.



Inputs Yield Outputs

Figure F10.1 Functional Organization for Quality.

ModelFool proof product design

MethodsFool proof processes

MaterialsConforming to all requirements:

Quality, quantity, handiling Packaging,preservation etc.

MachinesCapable processes

MeasurementsInspection and allied test errors

within harmless limits

MaintenanceAvailability of all

Equipments and facilities

ManEntire Human resource in stateof self control, empowered and

committed

MonitoringAdequate audit system

Deploysquality

functionleading to

Conception of optimalproduct and processstandards and their

conformance, as inputs

Products and services ofquality, that delights, inright quantity, at rightplace, price and time.

Appearance

Appealing

smoothness

etc

Ass

ess

wan

ts o

f cu

stom

er

Ass

ess

wan

ts o

f cu

stom

er

(Iterative process)

Concentrate to bridge this gap

Functional

Satisfactoryperformance

Reliability

Economic

useful life

Plan

Quality

Fitness

for use

Plan

Do

Do

IMBIBING QUALITY AS A WAY OF LIFE THROUGH ITS ABC 119


11IMBIBING QUALITY AS A WAY

OF LIFE THROUGH ITS ABC

An attempt has been made to enunciate basic tenets of quality and enumerate indispensableactivities to put mission of quality culture for continuous improvement in place by usingalphabets A to Z. Their accomplishment tends to attain leadership in one�s field of activity.

A: ASSURANCE OF QUALITY IS TRULY POSSIBLE ONLY BY THE PRODUCER

Though the quality of sweets is enjoyed and judged by its takers, yet it is its maker alone,who knows its contents and the care taken during its preparation. Therefore, he alone is in aposition to assure or guaranty the purity of its contents, its nutritive value, its harmlessness,shelf life and usefulness. Only the producer, the task performer or the server can build qualityinto the product manufactured or service rendered by him. He is knowledgeable about thedegree of its fitness for use. He is the sole valid competent person to assure its real worth, noone else can. Since the knowledge of the former is primary and complete, while that of anybody else is at best secondary and incomplete and so will be his assurance howsoever muchwell intended.

In early twenties, the link. between the producer and the consumer was direct or almostdirect. The former understood the needs of the latter and took adequate care of it, for his ownsurvival.

By mid twenties, a gap developed between the two, with the incoming of chain ofproduction, inspection and marketing.

By late twenties, this gap widened further, by the augmentation of the chain by marketresearch, design, process technology, packing, shipping and servicing teams.

Accordingly the concept of �Assurance� by the supplier or the producer broadened toinclude every member of the team as the ultimate task performer. Each one needs to knowhis optimal role precisely and perform exactly the same and also remain updated. Each one'srole is unassailable.

Any amount of audit or accredition cannot substitute the confidence that the primary taskperformer can exude. The audit is sample based. It is prone to both sampling and non-sampling errors. Their contributions can be estimated, reduced to harmless levels andcontrolled economically. These are rarely assessed. Often these are substantial. Theirreduction, standardization and control go by default. It is more so, with respect to audits.



In this context, it will only be doing justice to ourselves and the society at large, if we getinto the habit of doing self introspection for reforming ourselves rather than criticizing orfinding fault with others for their correction. We need to fall in line with the dictum �charitybegins at home�, the law of nature, accepted universally. This is the proven dependable pathfor stable and steady continuous improvement leading to the, most sought after peace andprosperity.

ASSURANCE OF QUALITY IS ANTICIPATION AND PREVENTION,OF ADVERSE EVENTS BY THE TASK PERFORMER'S ACTION.

B: BUSINESS STRATEGY OUGHT TO BE QUALITY FIRST

Quality is multidesciplined, multi pronged and multi edged sharpest weapon to minimize thecosts and maximize the returns. It gallops profits for economic growth to combat inflation andunemployment, the twin chronic problems and thus ushers an era of prosperity. Quality is anall weather friend; be it era of scarcity or plenty, monopoly or competition, recession orinflation, war or peace, drought or floods; since it operates through optimal utilization ofresources or minimizing if not avoiding waste in all its forms, thus reducing the costs.

Besides being all weather friend, quality is friend of �one and all� of every segment of thesociety, be it marketing, designing, producing, servicing or consuming or whatever. In the longrun, and that is what matters, what is not in the interest of one cannot be in the interest of thesociety.

Hence quality first, not quantity.BEST IS NOT GOOD ENOUGH FOR OUR EXISTENCE,CONTINUOUS IMPROVE-MENT NEEDS SUSTENANCE.

C: CUSTOMER FOCUS IS THE LIFE LINE OF QUALITY MISSION

Mahatma Gandhi had said:

A customer is the most important visitor on our premises.He is not dependent on us. We are dependent on him.He is not an interruption on our work. He is the purpose of it.He is not an outsider on our business. He is a part of it.We are not doing him a favour by serving him.He is doing us a favour by giving us an opportunity to do so.

We often hear that the customer is king or God. Do we ever revere or venerate him as suchor only dramatize for their make belief?

A dissatisfied customer rarely complains, dissuades many others and himself switchesover. Every single complaint compounds to over one thousand alarms. Is the complaint everviewed as such a monster?

No wonder, ISO: 9000 model for quality system starts with the assessment of the �Needs�of the customer, urges its compliance and proceeds for evaluation of his satisfaction to focuson the gap to be bridged for its enhancement. All quality models worth their name need to doso with emphasis on continuous improvement.



The concept of customer has rightly developed from ultimate user, to include next operatorin the chain and subsequently to everyone else, other than self. The self may be a customer ofevery one else.

The love for the other must be genuine. It should be selfless like that of a mother for herchild. She anticipates the needs and leaves nothing to chance to fulfill these. She puts herheart and soul into her mission. She sacrifices every thing at her command, offers no excusesand makes no demands, yet achieves her mission of fulfilling even the implied needs of thechild. Can such a love be a subject matter of assessment, audit or conformance?

The key is to make novel use of the resources available in fulfillment of the task ormission on hand, and not to cry about lack of any of these or offer lame excuses. One needs toremember and practice the dictum: �A satisfied customer is the supplier�s best unpaidsalesman�. Can a certificate of conformance ever match the role of a satisfied customer?

CUSTOMER DESERVES SELFLESS LOVE, THAT OF A MOTHER, BESTOW IT.

D: DELIVERY IS A PERTINENT PARAMETER OF QUALITY

Delivery on time, is crucial to all types of operations. Each organization is both a supplier anda customer. Thus quality system is universally applicable. Initially the concept of just in timewas focused on the delivery schedules only. A delayed delivery would upset productionschedules, necessitating sudden changes on the shop floor. These add to the unproductiveactivities of human resource, which do not have any value addition. These also increase theidle time of equipments or machines. To counter this menace, an easy way of buildinginventory is resorted to. This also drastically adds to the costs. Together, these cripple theeconomy, some times even beyond salvation. What use is the quality medicine that is sure tocure, if it is delivered after the patient is no more? Optimization tools for scheduling inventory,production and maintenance, etc; are available to meet a variety of situations.

The concept of �just in time� has been rightly extended to the execution of every activity,in the chain, as per optimal plan exactly - neither early, nor late, thus pegging the costs to becompetitive. Just in time leads to or is lead by TQM, if only the attempt either way is wholehearted.

DELAYS IMPAIR BEYOND REDEMPTION, BETTER ENSURE ITS ABSTENTION.

E: ECONOMICS OF QUALITY IS A MYSTERIOUS MULTIPLIER

Lack of financial resources is voiced as the root cause for POOR INDIA failing to make aniche in the quality world, to justify the need for subsidy and commercial protections from thegovernment for SURVIVAL. Host of other convenient, apparently logical and convincingarguments are advanced. Often these are only plausible excuses. Such measures only putsurvival at stake.

The best measure of economics of quality, easy to perceive, is the loss, both tangible andintangible, that the lack of quality imparts to the society. For example, a failure of a microcircuit breaker costing less than one hundred rupees may cause fire, resulting in loss of livesand property worth crores. A failure of a component of a locomotive may cause many fatalitiesand huge loss of equipment. Besides, delay in its restoration may jeoparadize availability ofvital perishable goods in transit, with accompanying stakes for the life of the people. A failure



of a social or business call may cause agony or loss of lifetime opportunity. A surgical errormay cripple the patient and his family for life. A poorly washed linen may cause painfulinfection that may prove fatal or the treatment of which may cost fortune. Soldiers' life, warand freedom may be lost if shoe laces are weak. A failure in power generation or supply chainmay bring the whole city or the entire nation to halt. Such examples of small errors ordeficiencies in quality that make or mar are in galore.

Every business calls for some investment. The returns on investment in quality are higherthan in any other form devoid of quality. These may vary from ten to forty times of theinvestment or even more. None of the countries having recognition in quality used subsidy asa stepping stone, nor the countries counting on subsidy have made a mark. Good qualityalways has a premium in the market; be it local, regional, national or international. Poorquality goes a begging sooner than later. A competent designer will give a better design at alower cost to meet the specific needs of the customer and a competent processor will processit competitively on time to capture the market. Today, the world is shrinking and we areperforce a player in the global market.

Hence, it is a myth that quality costs more. In fact, its overall life cycle cost per unit ofuseful life is always lower. Experiences of successful organizations corroborate this reality.The rising inflation and unemployment and falling value of rupee are enough evidence of poorquality culture. The cost of poor quality, in the form of losses imparted, is colossal. Any poorcountry would have perished by now. Only India, endowed with the richest natural resourcesin the world by the almighty, could and has survived so far. Indians are among the poorest bydint of their karma or complacency. In contrast, though Japan is among the poorest in respectof its natural resources, the Japanese are among the richest by dint of change in theirmanagement style and work culture. India has the potential to be the world leader. Japanesequality and economy were in shambles in early twenties. If they could change, so can we andfaster too. We must stop bragging about our plans and start working on these. Theglobalization phenomena would provide the needed impetus.

Statistically planned experiments lead to valid conclusions at reduced costs. These havebeen massively used by the developed countries to arrive at optimal, fool proof product andprocess designs termed off line quality control. These are strictly adhered to duringmanufacture termed on line quality control to deliver assured quality. These optimal levels bebetter adapted as product standards, usually national and process standards, usually plant.

The most frequently used designs are popularly called orthogonal arrays. The relevanttheory was provided by Dr C R Rao, as early as 1947. He is the recipient of many Nationaland International Awards and is popularly known as the best living statistician in the world.By introducing linear graphs and associated tables, Dr. G Taguchi has made the art ofdesigning experiments, to meet the specific needs, easy. Its potential has yet to be realized inIndia. These methods provide both rapidity and economy. Experiments have made thejourney, from stone age to space age, possible. Major era used classical methods ofexperiment. Statistically planned experiments have the potential to accelerate thedevelopment many folds.

EXPERIMENT FOR ADVANCEMENT.



F: FIXING THE PROBLEM AND NOT THE BLAME,IS THE NAME OF THE QUALITY GAME

Who created the non-conformity, is unimportant. How it happened is vital. His first handknowledge as a spot witness is an asset for the working group popularly called Quality Circlesor Quality Improvement Teams or Zee Dee (zero defect) Committees or six sigma groups insearch of the remedy, through leads provided during their brain storming sessions.

Brain storming sessions provide wonderful opportunity to the team to work successfully tofix the problem and not dissipate energy in blaming, witch-hunting or fighting among selves.Anyone can commit an error. It is futile to trace to pin him down. There is need to unite todevelop foolproof system and methodology and put these in place to forestall the recurrence ofthe problem. In this manner we march forward with conviction, smash every fresh problemencountered one by one and sustain growth rate of improvement. Statistics provides a kit ofversatile cost effective tools to win the quality game.

FIX THE PROBLEM, NOT THE BLAME.PARDON AND PARTICIPATION TO SOLVE IT, ARE ONLY NATURAL TO A

SUCCESSFUL LEADER.

G: GOOD INTENTIONS, PIOUS ETHICS AND NOBLE MORALS CATALYZESUCCESS

Just as healthy cultivation practices catalyse germination of seeds to maturity, in thepresence of right doses of fertilizers, environments, pesticides, and water; good intentions,pious ethics and noble deeds boost quality efforts in the presence of sound systems and aptmethodology. The latter are not sufficient by themselves. These do not substitute theintentions, deeds, honesty, and integrity of the human resource. However these do influencethe outcome of his passion and efforts to the hilt.

Global competition, in the offing, will generate the will to kill (excel). Lack of it shallnegate the growth to devastation.

GOAL OF QUALITY IS PRAGAMATIC, DYNAMIC AND RISING.

H: HUMAN RESOURCE IS VITAL ELEMENT OF QUALITY

All inputs of quality like model design, material, machine equipment including cutters, fixturesand jigs, etc., method process, measuring devices instruments and testing equipment,maintenance total, marketing and management systems including monitoring are inanimate.In the event of an occurrence of a fault, the only animate competent and committed humanresource can intervene timely to detect and correct the unwanted adverse change. This actprevents errors and constitutes genuine quality control. Thus indispensable human resourcedeserves parental development and sprucing up only. Gift it.

HUMANITY IS BEING HUMANE TO HUMANS.RESPECTING HUMANITY IS RESPECTING SELF.



I: INSPECTION IS NOT QUALITY CONTROL.YET, IT IS ITS INDISPENSABLE TOOL

Inspection to detect non-conformities to identify conforming items and documentation ofnature, severity and cause of nonconformity or its source is synonymous with postmortemreport of the dead. It cannot remove the fault or revive the dead. A faulty item is worse thana corpse. It can cause many fatalities. Does this event attract that concern?

On the contrary, self inspection by the operator, the task performer, of each and everyactivity in his domain consisting of inputs, processes and outputs, to adhere to the prescribedoptimal regime is the only sure way to build quality into one�s performance. This type ofinspection is eternal vigilance by the task performer. It does not cost even a single extra paisa.Yet, it is the price one has to pay. If there is any exception to this, it is not only worth it, butit shall prove the rule. A committed vigilant worker only can assure the quality with pride.

Direct cost of inspection consists of instrumentation, competent skilled human resourceand allied facilities. Indirect cost includes that of inevitable associated activities like additionalhandling, storage, holdups, delays and rework as a part of damage control exercise. Togetherthese can be colossal and even prohibitive. These make a difference between success andfailure.

The results of inspection, form the basis of decision making. Therefore assure the qualityof inspection. It should be compatible with the requirements of the situation, neither too goodnor too bad and cost effective too. Optimally calibrated and maintained apt testing facilitieswith documented proven standard procedures, for use in the hands of competent humanresource, are deemed assets to befit this need. Inspection errors, both sampling and non-sampling, like audit errors need to be assessed, reduced to harmless levels and containedinnovatively. Statistical analysis of inspection data assess contributions from various sourcesto guide determination of strategic priorities for needful actions. Even simple decisions,whether to manage with sampling inspection or one hundred percent inspection or both,taken rationally are known to have contributed to substantial savings.

Culturally, the domestic kitchen is taken care of by housewives. They are not exposed toformal course in an academic institution on management of kitchen. They inherit the art ofcooking and kitchen management by their ancestors. They provide the finest example of usingself inspection as a tool to prevent a failure. They serve food to the family members to theirtaste without waste. They do not have a formal manual on system or procedures, yet theyrightly and richly qualify for conformance to the requirements of I S O : 9000. They areaccustomed to follow the right regime as a habit, so much so, that no conscious effort seems tobe at work.

INSPECTION IS NECESSARY DEVIL, TAME IT, TO SLAVE IT.

J: JOY OF QUALITY IS ALL PERVASIVE

Joy of delivered quality is not limited to the customers alone, who are the main focus andbeneficiary. Every constituent of every segments of the organization is joyful with pride forcontributing directly or indirectly to the creation of such an item for the customer. They alsoshare it with their families. Similarly, the suppliers and the marketing agencies includingretailers without whom the mission would have been incomplete are party to this joy. Thus,the society as a whole is the beneficiary. It brings happiness and prosperity to all.



Monetary reward when shared by or distributed among more and more people getsreduced on percapita basis. Likewise the impact of sorrow also reduces when shared amongfriends, family members and will wishers. On the contrary the joy multiplies when shared.

JOY OF QUALITY IS UNIVERSAL. IT DELIGHTS ONE AND ALL.

K: KARMA, AND NOT THE SERMON, IS THE KEY FOR QUALITY PROGRESS

The Concise Oxford Dictionary defines karma as sum of person�s actions in one of hissuccessive states of existence, viewed as deciding his fate in the next. Further, the law of karmais the law of cause and effect. This law is scientific. It consists of two universal laws. First, aswe think so we become. Second, as we sow, so shall we reap. All our thoughts, words, deeds,emotions, feelings and wishes are seeds sown in the field of life. In due course of time theseseeds germinate, grow and bear fruits. Some of these are sweet and others bitter. All these tobe eaten by us, none else. The effect of some causes is felt immediately and that of someothers later. For example, the effect of over eating is felt soon after, while that of telling a liemay take some time. Remember every cause has an effect and every effect a cause. We need tobe judicious in choosing causative factors and sustaining these at proper levels to get desiredeffect or result.

The elements of karma are the steps of the ladder to reach the rising goals of quality.Actions and deeds need to follow the piously thought out plans like a shadow. No gaps,because time is money and the delays can dodge the mission just as proverbially as Many aslip between a cup and a lip. Time is instantly perishable commodity. Catch it from theforelock. Pious thoughts that are documented, say, as a Policy and preserved for referenceonly to be produced on demand for special occasions like audits are not sufficient. It is justlike having access to facilities treasured by the family for rituals on festive occasions only, tobe preserved again and forgotten till the next occasion.

Quality is not an act of chance. It is the result of concious efforts to frame and execute agood plan. Next, these get done subconciously. To reach perfection, these should flow rightly,naturally and even unconciously.

KARMA, TRANSLATING WORDS INTO DEEDS, SPELLS MAGIC FOR SUCCESS.

L: LEADERSHIP FROM THE TOP

The top management must get passionately involved and lead the noble mission of excelling,so much so, that the rest feel their omnipresence. It should take no chance by leaving the jobto the hired consultants, presumably though with good intentions. The role of consultants astrainers in theory and practice of methods is to catalyze. The company has to develop its ownprocedures and systems compatible with its culture and put these in place. This mission mustinvolve the entire human resource or else it goes by default. This honourable task is neitherdelegate-able nor shareable.

Concurrently, the top management needs to avoid short cuts and short term policies.These are evidence of narrow mind and greed. The instant gains might benefit, at best, thisgeneration only. On the contrary, there is need to depict a perspective long term vision tobattle for ultimate victory for next generations, that is, our children, grand children and so onand on. Such should be the culture and tradition in place. Create a citadel that no one elsecan venture to encroach or destroy. Only parental care and support through examples and not



precepts can make it happen. We have demonstrated our ability to plan and execute longterm policies while discharging parental responsibility of developing our progeny at home. Weneed to extend this potential, beyond the four walls of our homes.

LONG TERM VISION IN SERVICE OF SOCIETY IS SURETY, THAT THE REWARDSBOOMERANG INTO PROSPERITY.

M: MOTIVATION FOR QUALITY

Motivation for quality emerges from awareness and knowledge of potential loss and harm thatthe lack of quality can cause to the society. By nature, one longs to project image ofhelpfulness. Any exception to this axiom, shall not lead one to commit sin of harming others.The philosophy; help ever, harm never is the key to happiness.

Thus, imparting knowledge of loss and harm likely to be caused by the deviation in anyparameter of product, process or service in one�s area of work is vital part of training tomotivate. Any other form of motivation, particularly direct monetary incentives linked withperformance is short lived. The salary is intended for performing to one�s best ability. Can thelacunas be ever made up by extra reward or bribe?

Unexpected honour, recognition and promotions that follow quality performance, withadditional responsibilities, accompanying status with allied benefits and above all, selfcontentment, though invisible, constitute long term gains and are most powerful motivators.This is reflected in the culture of the organizations. It is indexed by rare turnover ofemployees. People Join but do not quit, unless superannuated. Even their superannuation isdelayed. These cannot be bound by rigid rules of law.

MOTIVATIONS THAT RECOGNIZE NOVEL CONTRIBUTIONS INNOVATIVELYWITH AFFECTIONATE GRACE, RESPECT AND HONOUR ONLY SUSTAIN.

N: NURTURE QUALITY CULTURE

Actions speak louder than words. Therefore, all actions must depict true commitment andrespect for the enunciated quality policies, procedures and systems. Glamorous displays areshort lived and suicidal. Development of quality culture is a mission to embrace the entirehuman resource with customer focus as the central theme. It requires immense patience andstamina. Remember, even hurry takes time. Yet, there is no room for complacency.

The chalega culture, the culture of accepting or tolerating the status as it is in vogue, hasto go. Instead the culture of continuous improvement possible through chain of breakthroughstudies should replace it. Globalization will put the challenge for survival in place. India hasalways proved its mettle whenever challenged.

Audits, as a rule or courtesy are carried out on prior intimation and planning. All of usremember our school audits or inspection days. The school is bedecked cosmetically as a brideto be and other preparations are made to put the best foot forward to allure the audit orinspection team for the day. The next day, the school is back to square one. It is also rare,that an eminent school with strong moral, academic and sports culture steals the show or getsrecognized. Random audits do provide glimpses of the true culture in place. But these areforbidden. Bias does creep in. Unless every day can withstand the audit, without any specialpreparation for the day, whatsoever, the desired culture is not in place.

NEVER GIVE UP. MORE SO, THE MISSION OF IMPROVEMENT OF QUALITY.



O: OPERATOR IN �STATE OF SELF CONTROL� ONLY, CAN DELIVER, GOODSAND SERVICES OF QUALITY

Everyone does work and is earnest to excel. If one performs the right way, it boosts theeconomy; and if one does the wrong way, it only plunders the economy. The problem boilsdown to knowing and practising the right method of doing things. Quality is an art of doingthings better, in the sense of superior quality or lower costs or both, in relation to thecompetitor(s) at any given point of time or self over time.

This is possible only when every operator is in a state of self control. This is the only wayto assure quality through prevention of non-conformity. The operator is deemed to be in thestate of self control if he or she:

● is aware of the why of the quality needs. This implies that he or she has the requisiteknowledge of the damage, harm and loss the lack of any desired quality feature canimpart to the customer or the user in particular, and the society at large. In short, heor she knows the why of the product design.

● knows the how of the process, that is, what exactly needs to be done and theconsequences of not adhering to the stated regime.

● has adequate means to perform the task assigned.● has the knowledge and means to assess what is happening and its outcome, so that one

can notice any adverse deviation as soon as it occurs.● has the knowledge and means to correct the same.Once the operator is in the state of self control as stated above, doing things right instead

of wrong costs nothing extra. Yet, the former contributes to the profit and the latter to theloss.

For all practical purposes Company Wide Quality Control CWQC, Total Company WideQuality Control TCWQC, Statistical Quality Control SQC, Total Quality Control TQC, TotalQuality Management TQM; Total Integrated Quality Management TIQM, Synchronised andIntegrated Total Quality Management SITQM or host of other similar nomenclature like JIT,Standardisation, Six sigma, SPC, MBO or VE imply the same mission. All these routes lead tothe goal of prosperity provided these are practiced without leaving a gap and lip sympathy isgiven a go bye.

In this context, the term operator extends to include every member of the human resource.Every one needs to be trained, facilitated and empowered adequately to be able to performright to accomplish the mission.

ORGANIZE TO FACILITATE AND EMPOWER THE HUMAN RESOURCEADEQUATELY TO THE LEVEL OF SELF SUFFICIENCY, TO BE A WINNER.

P: PANACEA FOR SOCIOECONOMIC PROBLEMS LIES IN PRACTISING QUALITYAS A WAY OF LIFE

Practice of Quality, as a Way of Life is anticipating the problem and fixing it. In simplewords, it is an act of prevention. Experience gained during transition from stone to space agehas affirmed prevention is better than cure and stitch in time saves nine. The latter issuggestive of timely maintenance and that the returns on this are likely to be ten fold.

The true potential in India is many times over. In the eyes of the professionals from thedeveloped nations, our plans are globally the best, while their execution among the worst.



Because of this dichotomy between Plans and Actions, it should be surprising if problems donot persist. These have, for over 50 years. One single major index is devaluation of rupee byover 10 and 20 times in relation to dollar and yen respectively since independence.

PLAN OPTIMALLY, PERFORM ACCORDINGLY, REVIEW AND IMPROVE.

Q: QUESTION REPEATEDLY TO ACCESS THE ROOT CAUSE OF THE PROBLEM

Ask questions that prompt journey to the initial cause of the adverse effect that needs to beremoved. Ask Why again and again, till the root cause of the problem is reached. Its removalwill foolproof the process to yield Zero Nonconformity. Besides Why, there are six more friendsready to serve this noble cause. These are What, When, Where, Who, Whom and How. Avoidthe unfavourable events through fool proofing and cultivate the favourable ones to sustain theapt levels of the causative factors to advantage.

QUALITY FIRST, QUANTITY WILL FOLLOW. THE CONVERSE IS NOT TRUE.

R: RIGHT FIRST TIME AND EVERY TIME

Doing right first time is the most economical way to build quality into the product or service.This eliminates inspection, which is unproductive, nay, counter productive. It consumesvaluable Effort, Facilities and Time. By doing right first time, associated holdups, reworkand delays simply vanish.

The total savings on these counts, both direct and indirect augment power to compete andboost economy for substantial growth. The right way is unique. It is the optimal in the givensituation. Each one in state of self control is its prerequisite, for one to be able to do right allthe time.

RIGHT IS ALWAYS THE RIGHT CHOICE.GIVE IT THE FIRST PREFERENCE.

S: STATISTICAL CULTURE IS THE BACKBONE OF QUALITY CULTURE

Statistics is a science in search of truth. It serves all other sciences. It is master of none. Itdelves into the void to know the unknown, to traverse from uncertainty to certainty. It is thekey technology. It consists of formulating the problem rightly, gathering relevant andadequate data, aptly analyzing, validly concluding, making confirmatory trials andimplementing these to reap the expected benefits and to repeat the cycle to continue the chainof improvements to reach the goal. Prof. P C Mahalonobis FRS founder director of the IndianStatistical Institute said so. He was a renowned Physicist, Economist, Architect of 2nd 5yearPlan and Adviser to the Government of India.

Statistics provides means to estimate the deviations from the target, risks, errors orvariations and the losses arising from these with desired confidence. Further analysis of theseto gauge the contributions from various sources prioritises the approach to optimize. Statisticsprovides a kit of indispensable universal tools for feasible solutions to the problems. Beware,it's unscrupulous use casts shadow on its credibility.



ISO makes its use obligatory. Sorry, no option, better start wooing it to enjoy its beauty,support and enlightenment. Surely, it is dependable, fascinating, rewarding and trustworthyscience.

STATISTICS IS SIMPLE, QUICK AND COST EFFECTIVE TECHNIC.IT IS EASY TO LEARN, PRACTICE AND HARNESS TO CLICK.

T: TRAINING IS THE ESSENCE OF THE TOTAL QUALITY MISSION

Continuous training and retraining to update is the heart and soul of every quality activity. Itis not confined to anyone discipline. To dominate in the war on poor quality, training, not onlyin job methodology and human relations, but also in intertwined disciplines like benchmarking or goal setting, communication, cost analysis, decision making, environments,ergonomics, forecasting, health, house keeping, JIT; management by objectives, exception andflexibility; marketing; material management including storage, inventory and supplier's ratingand development; motivation, operations research including optimization, PM, PERT & CPM,programming, queueing, reliability, scheduling, safety, standardization, systems; statisticsincluding SPC, sampling and experimentation; terotechnology, time management, valueanalysis, waste control, and work sampling is essential. This list is not exhaustive. It ought tobe multidisciplinary and cross functional. Each one needs to be proficient in one's domain. Thefield of training extends even to the compounds of the customers and the suppliers andconstitutes an eternal loop.

Richness and accessibility of library facilities is a vital index of organization's love forgrowth and development of its human resource.

TOTAL INTEGRATED AND SYNCHRONISED MANAGEMENT FOR CONTINUOUSIMPROVEMENT, STARTS WITH TRAINING AND COHABITS WITH IT ETERNALLY.

U: ULTIMATE IN QUALITY IS EXCELLENCE

Excellent performance is possible only through pious thoughts and noble deeds. In otherwords quality is a way of life in service of mankind or society of which self, each one of us, isa part. One is a share holder in what one gives to the society. Do good, share its fruits. Onecannot miss one�s share in the outcome of errors made, whether intended or not. Thereforeavoid errors. We are responsible for not only, what we give to others or society, but what weourselves really are or what we receive. This is in tune with: As we sow, so shall we reap.

USHER AN ERA OF PROSPERITY AND HAPPINESS,BY CRAVING FOR AND CARVING EXCELLENCE.

V: VALUE ENGINEERING IS THE PERFECT KEY TO QUALITY PROBLEMS

Value engineering provides very valuable guidance at every step by provoking analyticalapproach to choose from among the alternatives available, on the basis of net value added byconsidering cost versus benefit including worth ratio. Thus it provides right direction at everystep. It renders timely help to avoid unintended pitfalls. Trust this friend. Do not spare him!

VALUE TRULY THE PROS AND CONS.LEAVE NOT THE DECISIONS TO CHANCE.



W: WASTAGE SIGNALS QUALITY DEFICIENCY

Any resource that is improperly harnessed or under-utilised or idle contributes to the waste.Any amount of waste in any form, anywhere at any time is unflinching evidence of deficiencyin design of quality system or its operation or both. In contrast, ZERO WASTE is an evidenceof an adequate quality system in place. It ensures prosperity through growth augmented byproductivity, quality, value, worth and yield.

Wastes are generally invisible. These look more like a part of the necessary process andoften go un-noticed. These offer precious opportunities for improvement. These opportunitiesget lost by default. An abrupt stock out in spite of excess store enhances production delays,work in progress and delivery schedule. Similarly, a sudden breakdown or a nonconformityobserved at despatch station, causes enormous loss. Adverse chain of reaction sets in motion.These create unsurmountable loss of customer�s goodwill and invite fire fighting. Yet all theseare taken in as normal strides. Surprisingly, the total loss is large enough to win or lose. Itsestimation is quick and easy. Calculate the ideal production, assuming every step was takenright first time. Compare it with the actual. The gap between the two is the waste. These arethe gems in the waiting to be culled. Its awareness is Quality Consciousness and pursuitQuality Control.

Indians are among the best Fire Fighters. The joy of successful fire fighting blinds one tothe loss from suspension of normal work in progress. It is not rare to see some one to createfire and extinguish it to earn promotion. Instead, the energy needs to be routed to plan andexecute prevention of fire. Wise use of resources can reduce waste to zero. Unique use of allfacilities at hand to succeed and not to find scape goat in the absence of any of these for thefailure is quality control. Life is not possible without problem! Make it work for the solution.Consider two contrasting approaches to identical real life situations.

Two couples were blessed with a blind child each. One family, accepted the situation asthe will of the God and showered well intended, all possible care on him. This pamperingmade him dependent on others. No activity, led to ill health. Doctors advised exercise. He wasto run water hand pump, even though the family had servants to attend to such errands. Thefather in the other family happened to be an eye specialist. Hoping against hopes, he operatedto the best of his ability. As expected, it did not succeed. The family assessed and exploited hisother talents. He matured into a reputed singer. So much so he got an assignment in AllIndia Radio. He gained job, delighted the society, gifted direct job to one person to help him inhis daily chores, contributed to makers of musical instruments and transport services. He lednormal life and did not invite mercy, sympathy or charity from others. Such incidents provideperspective of quality culture versus lack of it.

Let us recall other heroes. A blind worker in HMT Pinjore delivered higher productionand superior quality among his colleagues. Kerala Cashew Industry engaged blind personsagainst reservation. To the surprise of one and all, they produced more and free of brokenkernels too. Social bonds led them to marry among themselves. Blindness is not hereditary.They bore normal children and led normal family life, perhaps better. These heroes havedemonstrated that they are second to none. They did not abuse the reservation privilege.They only needed equality of right to opportunity. They did not want charity. They fulfilledsocial obligations with flying colours by creating job opportunities for others, similarlyhandicapped.

In contrast the well intended reservation of 10 percent for 10 years introduced soon afterindependence has boomeranged. So much so that it has stretched up to 50 percent and in



some states even up to 70 percent. Soon it may be one hundred percent! There is craze toaccess national assets and grab rights. The prime need is to, first fill the treasury byperforming duties before exercising one�s right to its resources. One is obliged to depositbefore availing right to withdraw. The gear needs reversal. The father of the nation and othernational leaders have been giving priority calls for doing our duties. We have yet to imbibethese.

Let us look at one more dimension of resource utilization. Everyone is gifted with largenumber of useful limbs. Do we make apt use of these to serve the society? Consider eyes,hands and legs for instance. Perhaps, it is a moot question, whether we use these to help theneedy or to hurt the innocent. Once again these elucidate the mindset or attitude to quality asa way of life.

WASTE IS A CRIME AGAINST SOCIETY AND SIN AGAINST LAW OF DIVINE.

X: X-RAY OF ONGOING ACTIVITY EXPOSES POTENTIAL FOR IMPROVEMENTAND PROPOSES FIRST STEP TO ACCESS IT

Statistical process control (SPC) charts and allied analytical tools of shop floor data,synonymous with the On Line Quality Control, are akin to the terminology of X-Ray and ECGin the medical vocabulary. These provide true picture of current status of the health of theprocess and its owner. Thus, these aid timely detection and correction of adverse changes. DrW A Shewhart pioneered SPC approach in his book Economic Control of Quality ofManufactured Products. Often, these lead to identification of sources which if acted upon aptlyresult in reduction of variation and hence, improved intrinsic process capability. Pursuedsteadfastly, the process capability may approach half of the specified tolerance range. Theprocess capability index then equals two. This satisfies one of the two conditions of latest fadof six sigma status. The other condition is, that the process average of the controlled process,should not deviate from the target, by more than one-eighth of the tolerance, on either side.This, assures non-conformance of 3.4 ppm. If however, the average of the controlled processequals the target perfectly, the process assures non-conformance of only 0.1 ppm. One candetermine the most economic target, if the tolerance is one sided or the losses from over sizeand under size are unequal.

This wonderful inexpensive proven tool has yet to be exploited to its full potential. Themain reason for this lapse is that the relevant information on changes taking place in thelikely causes of variation, documented in the cause and effect diagram, are not recorded asand when they occur or the recording is inadequate. Thus the vital data, the analysis of whichcould provide the solution, is missing. The culture of recording these vital facts needs to beimbibed. There is no time to lose.

X-RAY AND ECG REPORTS VIZ STATISTICAL PRESENTATION OF DATA, AIDJOURNEY FROM SYMPTOMS THROUGH DIAGNOSIS TO CAUSE AND REMEDY.

Y: YIELD IS AN UNAMBIGUOUS INDEX OF QUALITY STATUS

Productivity, recovery or yield in the present context, implies consumption of any resource ofconcern per unit of Good or Conforming output. If the optimal capable system is in place, thewaste of each resource will be minimal and the yield maximal. The cost of production shall bethe lowest possible and this in turn shall make the enterpriser a leader in the market.



The ratios of the resources consumed to the conforming output will be smallest possibleand their respective contributions to the production cost too shall be minimal. These ratios ortheir inverses are obviously apt indices of the quality status in place. The word RESOURCEhas been used in broad sense to include tangible, intangible, direct and indirect inputs such asmachine hour, man hour, energy and time. Time is invisible and instantly perishable inputand hence the most precious. It deserves prime consideration continually.

YEARN FOR QUALITY, TO EARN PLENTY.

Z: ZERO DEVIATION FROM THE TARGET AND ZERO NONCONFORMITYCONSTITUTE THE IDEAL QUALITY THAT DELIGHTS THE CUSTOMER

It is not that Zero Variation and Zero Nonconformity are not possible. Where there is a will,there is a way. These are achievable and more often than not, these are achieved too. We failto achieve when we compromise.

We never miss a flight or a train. We are never late. The consequences of being late guideour actions. We are late to our work places, because we compromise. Once again theconsequences of being late to work weigh on our mind. This double standard is the bane forquality goals. Never the less, the achievements are ppm when it comes to missing a flight,collecting a lunch box from a customer�s residence and delivering the same at his work place,preparing a salary bill or a meal.

However, the wonder of wonders is that, as soon as these are achieved, the expectationsof the demanding customer rise instantly, leading to new challenges and targets, makingquality improvement an eternal phenomenal game. Beware! There is no escape and no time tolose. With faith in the Almighty and confidence in self, we better start the journey at once tothe best of our ability and integrity. Integrate and synchronize all the activities appropriatelyfor desired results.

ZERO TO HERO THROUGH DEDICATION, ETHICS, INTEGRITY ANDTECHNOLOGY.

It would be befitting to conclude with Swami Vivekananda�s sermon:

DIVINITY GETS UNFOLDED INSIDE THE MIND OF HUMAN WHEN HE DOESIMPROVEMENT. HUMAN SOCIETY WILL BECOME DIVINE, IF ONLY ALL OF US CANACQUIRE THE HABIT OF DOING IMPROVEMENTS.

APTITUDE TEST 133


12APTITUDE TEST

12.1 OBJECTIVE

The test has been compiled to enable the reader of the book and user of the tools to assessone�s proficiency in understanding basic concepts of quality tenets and tools dealt with formaking improvements through their application. This also enables one to identify the weakareas and the quantum gap that needs to be bridged. It may be desirable to time one self.

Each statement has four options. Read carefully and Circle, as the CAPITAL alphabetassociated with the most appropriate alternative from among those listed. Correct, un-attempted and incorrect responses invite scores of 4, 0 and -1 respectively.

12.2 TEST

1. Quality means,

A. Customer satisfaction.B. Produce of proven product and process designs.C. Product conforming to mid design specifications with least variation around it.

D. All the above.

2. Quality control means,

A. Prevention of nonconformity.B. Detection and correction of adverse change.C. Anticipating harmful effects of changes in inputs beyond control assessed through

appropriate monitoring and manoeuvring the rest controllable to undo the same.D. All the above.

3. Cause and Effect Diagram presents diagrammatically the causes, underappropriate heads and subheads which influence the effect of the problem understudy. These causes are as,

A. arrived at a brain storming session among all those associated directly and indirectly.B. listed in technology literature.C. dictated by the technical chief.D. short listed.



4. Check sheets are necessary for,

A. all the causes listed in the corresponding Cause & Effect Diagram.B. all the causes impacting all the appropriate indices of the effect.C. appropriate selected list from among A and B above. This may exclude such causes

and effects which are definitely known to be constant.D. none of the above.

5. Pareto analysis is a technique to help identify the vital few,

A. causes of nonconformities based on its frequency of occurrence.B. causes of nonconformities based on the losses it imparts.C. causes of nonconformities arising from negligence.D. sources of zero nonconformity.

6. Stratification is an art of collecting data such that,

A. all details are available.B. the guilty personnel can be identified,C. the deficient inputs can be identified through grouping the data in a manner that

inter group variation is large and intra group small.D. all the above.

7. Fundamental principles of stratification are,

A. the corners of excellence in all places of work.B. the availability of the data on the likely sources of deviations that are true and enable

traceability.C. consistent results that are repeatable under conditions of homogeneous inputs.D. all the above.

8. Scatter diagram is a tool to make a primary evaluation of the interdependence oftwo variables to assess,

A. the model of interdependence.B. the strength of interdependence.C. likely potential gain from the control of the independent variable.D. all the above.

9. Scatter diagram needs to be drawn for two variables only when,

A. relationship between the two is known to be linear.B. nature and strength of relationship needs to be assessed.C. two by two tabular or matrix data are seen.D. none of the above.

10. Histogram is a diagram that exhibits,

A. the value on x-axis and its frequency on y-axis.B. causes on x-axis and the loss on y-axis.C. both A and B.D. none of the above.

APTITUDE TEST 135


11. Histogram is likely to show,

A. central tendency and spread.B. likely hood or otherwise of the capability of the process, given the specifications.C. the extent of nonconforming units if the specifications are superimposed.D. all the above.

12. Run chart shows,

A. the parameter of interest on y-axis and the sequence of production on x-axis.B. the sequence of production on y-axis and the parameter of interest on x-axis.C. both the above.D. none of the above.

13. The prime responsibility for quality lies with the,

A. Top management personnel.B. Primary workers.C. Quality department personnel.D. All the above.

14. Careful examination of run chart and histogram indicate periods influenced by,

A. chance causes only.B. assignable causes only.C. both A and B, to aid determination of favourable levels of the factors for attempting

improvement.D. none of the above.

15. For achieving good quality, the most important inputs are,

A. fool proof product and process designs.B. competent, trained and committed human resource.C. raw materials of good quality.D. capable and well maintained machines including accessories.

16. Run chart as a tool is useful for,

A. only process control.B. only process improvement.C. both A and B.D. none of the above.

17. For effective product control, it is necessary to,

A. assess the needs of the target customer segment.B. develop fool proof product and process designs.C. develop adequate systems for all associated activities and put these in place in the

hands of competent empowered human resource.D. perform all the above functions.



18. Better quality is always associated with,

A. high returns.B. high cost.C. low costD. none of the above.

19. Process control is achievable through,

A. First off inspection.B. Patrol inspection.C. Last off inspection.D. all the above.

20. Pareto analysis is done through graphical presentation in the form of a barchart and a cumulative percent curve with,

A. percent contribution on x-axis and causes on y-axis.B. causes on x-axis and their contributions on y-axis.C. causes arranged in non-ascending order on x-axis and their percent loss contributed

on y-axis.D. all the above.

21. To ensure hazard free subassembly and assembly, it is necessary to,

A. control the process of the capable machines to the desired target.B. check all the components one hundred percent and use only those conforming to the

tolerances.C. use only the skilled operators, who are good at selective assembly and associated

rectification jobs.D. subcontract the job on piece rate basis.

22. Quality Management System and Methods aim at,

A. customer satisfaction.B. conforming processes, products and or services.C. conforming to legal and government regulations.D. all the above.

23. Check sheets are planned formats to record observations made by,

A. visual examination, measurement, gauging or test.B. visual examination only.C. go-no-go pair of gauges and measurement only.D. tests carried out on sophisticated equipment for critical parameters only.

24. Check sheets need to be designed to record data on,

A. critical parameters of vital products only.B. any appropriate indices of inputs and outputs of concern of any activity of interest.C. vital parameters of critical processes.D. none of the above.

APTITUDE TEST 137


25. A good check sheet should,

A. be easy and speedy to record and summarise.B. facilitate interpretation for planning and taking necessary action.C. both A and B above.D. none of the above.

26. Relationship between two variables x and y as seen from a scatter diagram issaid to be positive, if

A. y increases as x decreases.B. x increases as y decreases.C. x increases as y increases.D. none of the above.

27. Relationship between two variables x and y as seen from a scatter diagram issaid to be negative, if

A. x and y increase simultaneously.B. y increases as x decreases and vice versa.C. x and y decrease simultaneously.D. none of the above.

28. While drawing a scatter diagram, it is advisable to represent,

A. dependent variable on y-axis and independent on x-axis.B. dependent variable on x-axis and independent on y-axis.C. both A and B.D. none of the above.

29. To improve visual appeal of a scatter diagram the scale for x and y variablesshould be such that,

A. the spread on x and y axes are as much equal as possible and fairly large for properperception.

B. it is large for x and small for y.C. it small for x and large for y.D. none of the above.

30. If a histogram of 50 consecutive units produced with all controllable inputparameters operating at constant level, shows a normal pattern, it may be inferredthat the process is in state of,

A. economic control.B. statistical control.C. functional control.D. objective control.



31. It is true that,

A. an histogram of any measurable parameter of any product shall always depict anormal pattern.

B. an histogram of any measurable process parameter shall depict a normal pattern.C. if the process is in a state of statistical control and quality parameter variable

(measurable) the resultant histogram shall reveal the normal pattern.D. if an histogram shows a normal pattern, it is necessary that the data have emerged

from a stable process.

32. If a scatter diagram shows no evidence of relationship between two variables yand x, while technologically a relationship is expected, it may be inferred that,

A. accuracy of data generated may not be sufficiently precise.B. either x or y or both are controlled too rigidly.C. y may be influenced more by some unsuspected factor other than the conjectured x.D. any one or any combination of all the above.

33. Quality control means inspection of,

A. all parameters at all stages one hundred percent.B. input and process parameters, preferably by the operator himself, to sustain these at

appropriate levels, as means to control the process for controlling the productparameters for product assurance, at appropriate intervals

C. only critical parameters of critical stages of critical operations only.D. every minute detail, very rigorously, by an independent agency.

34. For meaningful linear relationship, if n1, n2, n3 and n4 are the number of pointsin quadrants I, II, III and IV formed by the medians parallel to the axes x and yrespectively, on a scatter diagram,

A. n1 = n2 = n3 = n4B. (n1 + n3) and (n2 + n4) should differ by a wide margin.C. (n1 + n3) = (n2 + n4)D. (n1 + n2) and (n3 + n4) should differ by a wide margin.

35. The strength of a linear relationship as seen from a scatter diagram is judged by,

A. the slope of the line of best fit, the larger the better.B. spread around the line of best fit, the smaller the better.C. both A and B.D. none of the above.

36. The best line of fit, in the context of linear relationship between two variablesimplies,

A. equal number of points on its either side.B. least sum of squares of deviations between the observed values of y and its values as

read from the line for the corresponding values of x.C. least sum of deviations observed and predicted or estimated values.D. none of the above.

APTITUDE TEST 139


37. The relationship determined from the scatter diagram between two variables xand y is,

A. limited to observed range of x.B. limited to observed range of y.C. both A and B.D. valid universally for all values of x and y.

38. Any linear relationship needs to be fully exploited for economic gains by,

A. increasing or decreasing x according as the relation is positive or negative, indefinitelyto extreme values.

B. confirming the nature of relationship beyond the observed ranges, till it shows upconcave pit or convex peak for optimum exploitation.

C. increasing x and y simultaneously.D. increasing x and decreasing y alternately.

39. If x and y are linearly related then,

A. increase in x will always be economical, if x and y are related positively.B. decrease in x will always be economical, if x and y are related negatively.C. increase in both x and y, if there is no evidence of any meaningful relationship.D. economically viable decision will depend on the cost of increasing or decreasing x by a

unit and the corresponding gain likely by the improvement in y depending on theslope of the regression line.

40. Given the specification of y of a product parameter, the specification of x for theprocess parameter showing linear relationship can be obtained by reading value ofx, by projecting lines from the upper and lower specified values of y, parallel to xaxis and reading the values of x corresponding to their points of intersection with,

A. upper and lower lines parallel to the regression lines, providing for process errorsrespectively.

B. regression line, the line of best fit.C. lower and upper lines parallel to regression line providing for process errors

respectively.D. none of the above.

41. One of the key indices of the quality status of an organization is,

A. the growth rate of the market share.B. the market share.C. pending orders.D. loan advanced.

42. Quality has been defined as fitness for use. It emphasizes that it is the customerwho decides the degree to which fitness for use has been achieved. In this sense,

A. the customer is the next operator.B. the customer is the wholesaler, the retailer and the end user.C. the customers are the people around who are affected by it when the user is using the

product.D. all the above represent the customer.



43. Operator in the state of self control means that,

A. he knows the why of the product & what of the process he is performing and has themeans to do the same.

B. he has the means to know what is happening as also the knowledge & means tocorrect when necessary and is empowered to do so.

C. both A and B above.D. he has full authority to do anything in his area of work.

44. The cost of providing quality goods and services to the customer includes thecosts of prevention, appraisal and failure. If prevention cost is increased to augmentpreventive efforts then the other costs are very likely to,

A. remain the same.B. decrease.C. increase.D. increase or decrease such that the total cost will either remain the same or go up.

45. For the salutary effect of the application of the seven simple, quick and costeffective statistical tools, these must be used in the sequence in,

A. which these are published.B. which these are discussed.C. any combination and sequence best suited to the situation.D. none of the above.

46. While conducting brain storming session for drawing cause and effect diagram,the leader should,

A. not interfere in a fashion that might discourage others from contributing or sharingtheir experience.

B. encourage and invite every one to contribute.C. repeat exercise to arrive at total comprehensive picture.D. take care of all the above.

47. The utility of the run chart is enhanced if the,

A. operator is penalized for the errors.B. operator records all the changes occurring in the inputs and allied process parameters

listed in the cause and effect diagram and the details of actions (nature and quantum)taken to remedy the adverse event in his domain of work.

C. operator checks and records only when he receives a complaint from the owner of thenext process or any of the subsequent operations.

D. operator is paid on piece rate basis.

48. Quality improvement is synonym with,

A. continuous attempt to reduce deviation from the optimal target.B. continuous thrust to approach zero nonconformity.C. both A and B.D. none of the above.

APTITUDE TEST 141


49. The operator wise stratification of nonconformities is attempted to,

A. identify the operators, responsible for the errors and losses, to hold them accountablefor awarding appropriate punishment under the labour laws.

B. utilize their experience as first hand witness of the adverse situations that caused theerrors, for making improvements.

C. benefit from both the above.D. keep the operator under thumb, to discourage him from participating in union

activities.

50. Interpretation of the process data with the aids of the run chart andcorresponding histogram has potential to,

A. expose the scope for improvement.B. propose first step in the right direction to reduce variation.C. do both the above.D. none of the above.

12.3 KEY TO THE ANSWERS TO THE OBJECTIVE TEST

1. D 11. D 21. A 31. C 41. A2. D 12. A 22. A 32. D 42. D3. A 13. A 23. A 33. B 43. C4. C 14. C 24. B 34. B 44. B5. B 15. B 25. C 35. B 45. C6. C 16. C 26. C 36. B 46. D7. D 17. D 27. B 37. C 47. B8. D 18. A 28. A 38. B 48. C9. B 19. D 29. A 39. D 49. B

10. A 20. C 30. B 40. A 50. C



LIS

T O

F C

AS

E S

TU

DIE

S

Ref

eren

ceIn

du

stry

Gro

up

Pro

du

ctP

roce

ssIt

emA

ppro

ach

Pro

blem

/B

enef

it

5.51

aC

hem

ical

Van

aspa

tiS

olde

rin

g16

.5 K

g T

inS

trat

ific

atio

n b

yE

lim

inat

ion

of

sold

erin

gop

erat

or.

rew

ork

and

rein

spec

tion

.b

En

gin

eeri

ng

Tin

Insp

ecti

on16

.5 K

g T

inS

trat

ific

atio

n b

yE

lim

inat

ion

of

cust

omer

man

ufa

ctu

rein

spec

tor.

retu

rn d

ue

to l

eaka

ge.

cC

hem

ical

Un

ani

Fil

lin

g an

dS

um

mer

dri

nk

Str

atif

icat

ion

by

Ou

t pu

t do

ubl

ed.

phar

mac

euti

cal

labe

lin

gw

orke

r.d

Ele

ctri

cal

Oil

Ass

embl

yO

ilS

trat

ific

atio

n b

yE

lim

inat

ion

of

reas

sem

bly

En

gin

eeri

ng

tran

sfor

mer

tran

sfor

mer

grou

p of

wor

kers

.n

eces

sita

ted

from

lea

kage

.e

Exp

ort

insp

ecti

onS

ervi

ceIn

spec

tion

Man

yS

trat

ific

atio

n b

yJu

dici

ous

iden

tifi

cati

on o

fin

spec

tor

and

pert

inen

t co

mpl

ain

ts a

nd

cust

omer

.in

spec

tion

s fo

r re

ctif

icat

ion

of s

itu

atio

n.

5.52

aC

hem

ical

Ray

on g

rade

Ch

ippi

ng

ofB

ambo

oS

trat

ific

atio

n b

yE

nh

ance

men

t of

pro

duct

ion

,pu

lpB

ambo

osC

hip

sC

hip

pin

g m

ach

ines

.im

prov

emen

t of

qu

alit

y an

dre

duct

ion

in

con

sum

able

s.b

Ele

ctri

cal

Cap

acit

orF

illi

ng

and

Cap

acit

orS

trat

ific

atio

n b

yR

educ

tion

in

rew

ork

to a

lmos

tE

ngi

nee

rin

gse

alin

gpr

essi

ng

and

nil

.se

alin

g m

ach

ines

.

5.53

aT

exti

leC

loth

Wea

vin

gC

loth

Str

atif

icat

ion

by

Str

ateg

ic p

lann

ing

to a

ugm

ent

Loo

m S

hed

shif

ts.

prod

uct

ion

an

d pr

odu

ctiv

ity.

bE

nvi

ron

men

tall

yS

uch

as

Hea

tin

g et

c.A

dhes

ive

Str

atif

icat

ion

by

Red

uct

ion

in

non

con

form

itie

sro

om t

empe

-ad

hes

ive

and

shif

ts.

and

repr

oces

sin

g.ra

ture

sen

siti

vese

nsi

tive

au

toop

erat

ion

spa

rts

cE

ngi

nee

rin

gL

arge

var

iety

Mac

hin

ing

Man

yS

trat

ific

atio

n w

ith

Eli

min

atio

n o

f n

on-

resp

ect

to t

ime

toco

nfo

rmit

y, r

ewor

k an

d sc

rap.

opti

mis

e to

olgr

indi

ng

and

chan

ge i

nte

rval

.d

Ch

emic

alE

lect

ropl

ated

Ele

ctro

-E

lect

ropl

ated

Str

atif

icat

ion

wit

hR

edu

ctio

n i

n n

on-c

onfo

rmit

ypa

rts

plat

ing

com

pon

ents

resp

ect

to t

ime

toan

d re

wor

k.op

tim

ise

inte

rval

for

repl

ensh

ing

and

chan

gin

g pl

atin

gba

th s

olu

tion

.

LIST OF CASE STUDIES 143


5.54

aP

har

mac

euti

cal

Med

icin

esS

ervi

ce/

Med

icin

esG

eogr

aph

ical

En

han

cem

ent

of m

arke

tT

onic

etc

.m

arke

tin

g/st

rati

fica

tion

for

shar

e/sa

les.

sale

sea

ch p

rodu

ct.

bS

hoe

Sh

oes

Sh

oeS

hoe

sG

eogr

aph

ical

Incr

ease

of

dura

bili

ty.

man

ufa

ctu

rem

akin

gst

rati

fica

tion

of

each

typ

e of

non

-co

nfo

rmit

y.c

En

gin

eeri

ng

Inst

rum

ents

Ass

embl

yA

uto

met

erG

eogr

aph

ical

Red

uce

d re

wor

k an

d sc

rap.

wat

emet

erst

rati

fica

tion

for

Impr

oved

qu

alit

y. E

nh

ance

dw

atch

esid

enti

fy v

ital

prod

uct

ion

.co

mpo

nen

t an

d it

spa

ram

eter

.d

Ele

ctri

cal

Fan

sA

ssem

bly

Fan

Geo

grap

hic

alE

lim

inat

ion

of

non

-val

ue

stra

tifi

cati

on b

yad

diti

on r

ewor

k du

e to

loca

tion

of

test

ing/

not

ion

al s

low

spe

ed.

insp

ecti

on o

n c

eili

ng.

eW

are

hou

seS

ervi

ceS

tora

ge l

oss

Foo

d gr

ain

sG

eogr

aph

ical

Bia

s, a

mon

g de

spat

ch a

nd

in t

ran

sit

stra

tifi

cati

on b

yre

ceip

t st

atio

ns

iden

tifi

ed,

desp

atch

ing

and

Impr

oved

con

ciou

snes

sre

ceiv

ing

stat

ion

sre

sult

ing

in r

edu

ced

loss

es.

to i

den

tify

rou

tes

pron

e to

mor

etr

ansi

t lo

ss.

fH

ospi

tal

and

Ser

vice

Was

hin

gL

inen

Geo

grap

hic

alW

ash

ing

effi

cien

cy i

mpr

oved

Hot

els

Lau

ndr

yst

rati

fica

tion

by

in t

erm

s of

qu

alit

y an

d co

st.

sou

rce/

wor

k/u

sage

area

of

lin

en.

gM

etal

lurg

yC

asti

ngs

,C

asti

ng,

Cas

tin

gs,

Geo

grap

hic

alR

edu

ced

non

-con

form

ity.

forg

ings

,fo

rgin

g,fo

rgin

gs,

stra

tifi

cati

on b

yfa

bric

atio

ns

fabr

icat

ing

fabr

icat

ion

slo

cati

on f

or e

ach

non

-con

form

ity.

hP

har

mac

euti

cals

Tab

lets

Tab

let

Tab

let

Geo

grap

hic

alH

oppe

r fe

ed c

ontr

ol i

nte

rval

mak

ing

stra

tifi

cati

on b

yst

anda

rdis

ed.

Hom

ogen

eity

leve

l of

mix

ture

of i

ngr

edie

nts

of

tabl

et a

nd

in h

oppe

r.it

s w

eigh

t im

prov

ed.

Ref

eren

ceIn

du

stry

Gro

up

Pro

du

ctP

roce

ssIt

emA

ppro

ach

Pro

blem

/B

enef

it



Ref

eren

ceIn

du

stry

Gro

up

Pro

du

ctP

roce

ssIt

emA

ppro

ach

Pro

blem

/B

enef

it

5.5

aC

hem

ical

Hyd

roge

nat

edH

ydro

gen

atio

nV

anas

pati

Str

atif

icat

ion

by

Rep

roce

ssin

g el

imin

ated

.oi

lsu

ppli

er o

f oi

l.b

Ele

ctri

cal

Fan

Pai

nti

ng

Bla

des

Str

atif

icat

ion

by

Pai

nti

ng

qual

ity

in t

erm

s of

can

opy

and

bran

d of

pai

nt.

har

dnes

s an

d gl

ossi

nes

sm

otor

bod

yim

prov

ed.

Cos

t re

duce

d.c

Ch

emic

alR

ayon

gra

deC

hip

pin

gC

hip

sS

trat

ific

atio

n b

yC

hip

pin

g ou

tpu

t in

crea

sed

pulp

bran

ds o

f V

-bel

tsan

d co

st r

edu

ced.

for

thei

r u

sefu

l li

fe.

dE

lect

rica

lF

ans

Ser

vice

Fan

Str

atif

icat

ion

by

Red

uct

ion

in

cu

stom

er(m

arke

tin

g/de

aler

s (c

ust

omer

s).

com

plai

nts

an

d re

turn

s.sa

les)

eP

har

mac

euti

cals

Bab

y to

nic

Bot

tle

Bab

y to

nic

Str

atif

icat

ion

by

Eli

min

atio

n o

f gl

ass

spec

ks,

was

hin

g,pr

oces

s st

age.

the

pres

ence

of

wh

ich

can

fill

ing

and

be f

atal

.se

alin

g

6.9

aJu

te m

ill

Jute

wov

enW

eavi

ng

Jute

wov

enS

catt

er d

iagr

amO

ptim

al s

peed

to

max

imis

e(l

oom

s)of

ou

tpu

t ve

rsu

sou

tpu

t.m

otor

spe

ed.

bA

uto

mob

ile

Veh

icle

Ru

nn

ing

Fu

elS

catt

er d

iagr

amO

ptim

al s

peed

to-

min

imis

e(c

onsu

mpt

ion

)of

fu

el c

onsu

med

fuel

con

sum

ptio

n a

nd

poll

u-

vers

us

run

nin

gti

on a

nd

enh

ance

lon

gevi

tysp

eed.

of v

ehic

le a

nd

safe

ty o

fpa

ssen

gers

.c

Ch

emic

alB

rew

ery

Bre

win

gB

eer

Sca

tter

dia

gram

Pro

vide

d cl

ue

for

nee

d or

of o

utp

ut

vers

us

oppo

rtu

nit

y fo

r be

tter

pro

cess

barl

ey c

onsu

med

.co

ntr

ol o

r in

com

pati

bili

ty o

fex

isti

ng

test

pro

cedu

res.

dE

ngi

nee

rin

gB

icyc

leP

edal

an

dN

on-f

ree

Sca

tter

dia

gram

Des

ign

in

adeq

uac

y id

enti

fied

.fr

ee w

hee

lpe

dal

of i

nte

rfer

ence

asse

mbl

ym

ovem

ent

vers

us

com

pon

ent

para

met

er.

eM

edic

alR

esea

rch

Pat

ien

t ca

re(T

o as

sess

)S

catt

er d

iagr

am o

fT

o de

cide

on

dru

g an

d it

sef

fect

of

dru

gef

fect

of

each

dru

gdo

sage

for

eff

ecti

ve c

ure

.ve

rsu

s it

s do

sage

.



Ref

eren

ceIn

du

stry

Gro

up

Pro

du

ctP

roce

ssIt

emA

ppro

ach

Pro

blem

/B

enef

it

fS

team

boi

ler

Ser

vice

Boi

lin

gS

team

Sca

tter

dia

gram

s of

Impr

oved

eff

icen

cies

of

stea

mat

tach

ed t

o(i

) S

team

gen

era-

gen

erat

ion

an

d co

nsu

mpt

ion

.h

otel

, h

ospi

tal,

ted

vers

us

wat

erva

nas

pati

, te

xtil

eco

nsu

med

wit

hch

emic

als

etc.

due

rega

rd t

ote

mpe

ratu

re.

(iii

) S

team

gen

e-ra

ted

vers

us

coal

(fu

el c

onsu

med

)(i

ii)

Ou

tpu

t ve

rsu

sst

eam

con

sum

ed.

gR

ubb

erS

hoc

k ab

sorb

ers

Cu

rin

gR

ubb

eris

edS

catt

er d

iagr

am o

fR

edu

ced

rew

ork.

man

ufa

ctu

refo

r ra

il w

agon

spl

ates

/dis

cs.

perc

ent

non

-co

nfo

rmit

y ve

rsu

sam

oun

t of

ru

bber

com

pou

nd.

hV

anas

pati

Hyd

roge

nat

edB

leac

hin

gH

ydro

gen

ated

Sca

tter

dia

gram

of

Red

uce

d re

proc

essi

ng

and

oil

oil

fin

al c

olor

ver

sus

enh

ance

d fi

nal

pro

duct

ion

.in

itia

l co

lor

of o

ilfo

r im

port

ed a

nd

indi

gen

ous

blea

-ch

ing

agen

tse

para

tely

.i

Hea

vy e

lect

rica

lT

urb

ine

Cu

rin

gR

otol

slo

tS

catt

er d

iagr

am o

fR

edu

ced

scra

p/re

wor

k.en

gin

eeri

ng

trou

ghpr

odu

ct p

aram

eter

vers

us

inpu

t or

proc

ess

para

met

er.

jC

hem

ical

Tit

aniu

m p

rodu

ctR

otor

y ki

lnT

itan

ium

Sca

tter

dia

gram

of

Impr

oved

qu

alit

y an

dan

d ce

men

top

erat

ion

diox

ide

prod

uct

ion

/pro

duct

enh

ance

d pr

odu

ctio

n.

para

met

er v

ersu

son

e by

on

e pa

ra-

met

er o

f fe

ed a

nd

proc

ess.



Ref

eren

ceIn

du

stry

Gro

up

Pro

du

ctP

roce

ssIt

emA

ppro

ach

Pro

blem

/B

enef

it

kR

ubb

erC

ycle

tyr

eD

raw

ing

orT

read

Sca

tter

dia

gram

of

En

d w

aste

an

d pr

oces

sm

anu

fact

ure

trea

d sh

apin

gw

eigh

t pe

r u

nit

insp

ecti

on r

edu

ced.

len

gth

ver

sus

adju

stm

ent

para

-m

eter

s li

ke h

eigh

t.

7.5

aE

ngi

nee

rin

gM

ach

ine

tool

sJo

b sh

opM

any

His

togr

am f

orR

edu

ced

scra

p an

dA

eron

auti

csea

ch d

imen

sion

asse

mbl

y H

ICC

UP

S.

of c

ompo

nen

tof

con

cern

.b

Ch

emic

alT

her

mos

fla

skB

low

ing

Gla

ss s

hel

lH

isto

gram

of

Str

eam

lin

ing

team

wor

kou

tpu

t (n

um

ber

wit

h p

oten

tial

to

augm

ent

prod

uce

d)pr

odu

ctio

n b

y 30

per

cen

tpe

r sh

ift.

per

un

it o

f ti

me.

cC

hem

ical

Org

anic

Boi

lin

gO

rgan

icH

isto

gram

of

pro-

com

pou

nd

duct

ion

per

hou

rov

er d

ays.

dE

ngi

nee

rin

gC

onde

nse

r pl

ate

Dri

llin

gH

ole

His

togr

am o

f di

a-R

edu

ced

rew

ork.

Im

prov

edE

lect

rica

lm

eter

s of

hol

es i

nco

nsi

sten

cy.

En

han

ced

a pl

ate.

prod

uct

ion

.e

Pla

stic

tex

tile

Wou

nd

bobb

ins

Win

din

gB

obbi

ns

His

togr

am o

f di

a-Im

prov

emen

t in

val

idit

y of

met

er o

f bo

bbin

s.in

spec

tion

an

d co

nsi

sten

cyof

dia

met

er o

f w

oun

dbo

bbin

s.f

En

gin

eeri

ng

Ser

vice

Insp

ecti

onC

ompo

nen

t fo

rH

isto

gram

s of

Insp

ecti

on s

tan

dard

ised

to

for

ampe

res

Tel

epri

nte

ram

pere

s ob

serv

edav

oid

wro

ng

deci

sion

s on

of a

on s

ame

com

po-

acce

ptan

ce o

r re

ject

ion

com

pon

ent

nen

ts b

y pr

odu

c-be

nef

itti

ng

prod

uce

r an

dti

on a

nd

insp

ecti

onco

nsu

mer

.pe

rson

nel

.g

En

gin

eeri

ng

Sh

avin

g pr

odu

ctS

tam

pin

g an

dS

hav

ing

His

togr

ams

for

Exp

lode

d m

yth

of

reje

ctio

nh

onin

gbl

ade

cen

tral

ity

and

due

to o

ld p

oor

capa

ble

hon

ing

angl

e.m

achi

nes.

Inv

este

men

t in

new

impo

rted

mac

hin

es a

void

ed.

hE

ngi

nee

rin

gA

uto

mot

ive

die

Die

cas

tin

gA

uto

mot

ive

His

togr

am o

fD

imen

sion

al r

ewor

k re

duce

d.ca

st c

ompo

nen

tpa

rtdi

men

sion

.



Ref

eren

ceIn

du

stry

Gro

up

Pro

du

ctP

roce

ssIt

emA

ppro

ach

Pro

blem

/B

enef

it

iE

ngi

nee

rin

gA

uto

mot

ive

An

nea

lin

gA

uto

mot

ive

His

togr

am o

fL

ot o

ne

hu

ndr

ed p

erce

nt

part

part

har

dnes

s.in

spec

tion

eli

min

ated

an

dre

wor

k re

duce

d.j

Pla

stic

Fib

reF

ilam

ent

Fib

reH

isto

gram

of

Was

te r

edu

ctio

n t

hro

ugh

form

atio

nbr

eaki

ng

load

appr

opri

ate

proc

ess

con

trol

for

each

gra

de.

kE

ngi

nee

rin

gE

lect

ropl

ated

Ele

ctro

-P

late

d pa

rts

His

togr

am o

f ba

thR

ewor

k re

duce

d, p

rodu

ctio

npa

rts

plat

ing

para

met

ers

enh

ance

d, u

nif

orm

ity

impr

oved

.l

Fer

tili

ser

Fer

tili

ser

Pac

kin

gF

erti

lise

rH

isto

gram

of

Air

con

diti

onin

g av

oide

d.w

eigh

t of

fil

led

bags

.

8.61

aE

ngi

nee

rin

gP

ress

sh

opD

raw

ing

Hem

isph

eric

alR

un

ch

art

ofS

crap

of

10 p

erce

nt

bow

ln

um

ber

non

-con

-el

imin

ated

.fo

rmin

g pe

r sa

mpl

e.b

En

gin

eeri

ng

Job

shop

Wel

din

gF

abri

cate

d jo

bR

un

ch

art

of n

on-

Ass

essm

ent

of q

ual

ity

and

con

form

itie

s.im

prov

emen

t fo

r th

e fi

rst

tim

e.c

En

gin

eeri

ng

Job

shop

Gas

cu

ttin

gF

abri

cate

d jo

bR

un

ch

art

ofD

emer

it s

core

im

prov

ed f

rom

dem

erit

sco

re.

35 t

o 5

per

job.

dE

ngi

nee

rin

gC

onde

nse

r pl

ate

Dri

llin

gC

onde

nse

rR

un

ch

art

for

Non

-con

form

ity

perc

ent

elec

tric

alpl

ate

hol

eh

ole

diam

eter

.re

duce

d fr

om 8

.0 t

o 0.

3pe

rcen

t.e

En

gin

eeri

ng

Job

shop

Mac

hin

ing

Tu

rbin

e bl

ades

Ru

n c

har

t (m

aste

r)N

on-c

onfo

rmit

ies

redu

ced

elec

tric

alof

non

-con

form

itie

s.fr

om o

ver

30 t

o u

nde

r0.

5 pe

rcen

t an

d pr

odu

ctio

nen

han

ced

fou

r fo

lds.

8.62

aC

hem

ical

Fer

tili

ser

Gra

indi

ng

Pu

lver

ised

Ru

n c

har

t fo

rV

aria

nce

red

uce

d to

on

ero

ck f

inen

ess

resi

due

on 1

00si

xtee

nth

.m

esh

.b

En

gin

eeri

ng

Tu

rbin

e bl

ades

Mac

hin

ing

Tu

rbin

e bl

ades

Ru

n c

har

t on

KR

Var

ian

ce r

edu

ced

to o

ne

elec

tric

aldi

men

sion

.tw

enty

fift

h.



Ref

eren

ceIn

du

stry

Gro

up

Pro

du

ctP

roce

ssIt

emA

ppro

ach

Pro

blem

/B

enef

it

cE

ngi

nee

rin

gA

uto

mot

ive

Mac

hin

ing

Au

tom

otiv

eR

un

ch

art

on(i

) P

oten

tial

to

redu

ceco

mpo

nen

tpa

rt d

iam

eter

diam

eter

.va

rian

ce t

o on

e te

nth

.(i

i) V

aria

nce

red

uce

d to

on

efo

urt

h.

dE

ngi

nee

rin

gR

otor

dis

cD

rill

ing

Rot

or d

isc

Ru

n c

har

t on

Sta

ndar

d de

viat

ion

redu

ced

toel

ectr

ical

dist

ance

in

dica

tive

one

twel

veth

(va

rian

ce b

yof

cen

tral

loc

atio

n.

1/14

4).

9.1

toH

eavy

ele

ctri

cal

Tu

rbin

eT

apin

g to

Tu

rbo

Use

d al

l th

e se

ven

Per

cen

t n

on-c

onfo

rmit

y9.

8en

gin

eeri

ng

curi

ng

gen

erat

or b

arto

ols.

redu

ced

from

abo

ut

5.0

to 0

.5pe

rcen

t.

10.2

Com

mu

nic

atio

nT

elep

hon

e ca

ble

Cab

leC

able

Des

ign

of

Fai

lure

du

e to

lea

kage

cabl

efo

rmat

ion

expe

rim

ent.

redu

ced

from

abo

ut

95 t

om

anu

fact

uri

ng

(im

preg

nat

ion

)ab

out

5 pe

rcen

t.10

.2In

sect

icid

esD

isin

fect

ant

For

mu

lati

onD

isin

fect

ant

Pro

cess

con

trol

.R

ewor

k re

duce

d fr

om t

hir

ty-

and

grin

din

gfi

ve t

o on

e pe

rcen

t.10

.2S

tate

tra

nsp

ort

Ser

vice

Ru

nn

ing

Mil

eage

Pro

cess

con

trol

.F

uel

con

sum

ptio

n r

edu

ced,

(Spe

ed a

nd

air

tyre

lif

e in

crea

sed.

pres

sure

).10

.2C

eram

icIn

sula

tor

Pu

g m

illi

ng

Insu

lato

rP

roce

ss c

ontr

ol.

Insu

lati

on p

rope

rty

impr

oved

and

baki

ng

to a

ccep

tabl

e le

vel

avoi

din

gsc

rap.

10.2

En

gin

eeri

ng

Com

pon

ent

Mac

hin

ing

Com

pon

ent

Pro

cess

con

trol

.R

educ

ed r

ewor

k on

com

pone

ntan

d as

sem

bly

and

hic

cup

free

ass

embl

y.

BIBLIOGRAPHY 149


BIBLIOGRAPHY

Ackoff, R.L. and Rivett, Patrick, Manager�s Guide to Operations Research. New York: John Wiley, 1963.

American Society for Quality Control; Wisconsin Quality Costs-What and How; Prepared by QualityCost-Cost Effectiveness Technical Committee�Wisconsin: The American Society for QualityControl, 1971.

American Society for Quality Control, Wisconsin Guide for Reducing Quality Costs; Report of theQuality Costs Technical Committee�Wisconsin: ASQC, 1977, p. 46.

American Society for Quality Control, Quality Motivation Technical Committee, Quality MotivationWorkbook�Wisconsin: ASQC Quality Pr., 1967, p. 126.

Asian Productivity Organization�Tokyo. (Industrial Engg. and Technology) Japan Quality ControlCircles; Quality Control Circle Case Studies 1972 (4th Reprint, 1984) p. 208.

Bajaria, Hans J.; ed. Quality Assurance, Methods, Management and Motivation. Michigan: Society ofManufacturing Engineers, 1981, p. 248.

Bandyopadhyay, Jayanta K. QS-9000 Handbook; A Guide to Registration and Audit. Florida: St. LuciePr., 1996, p. 244.

Barker, Thomas B. Engineering Quality by Design; Interpreting the Taguchi Approach, New York:Marcel Dekker, 1990 i, p. 250.

Bieda, John, Practical Product Assurance Management. Wisconsin: ASQC Quality Pr., 1999, p. 249.

Biswas, Suddhendu, Statistics of Quality Control; Sampling Inspection and Reliability. New Delhi: NewAge Intnl., 1997, p. 140.

Blazey, Mark L. Insights to excellence 1996; An Inside Look at the 1996 Baldrige Award Criteria.Wisconsin: ASQC Quality Pr., 1996.

Bossert, James L. Quality Function Development; A Practitioner�s Approach�Milwaukee: ASQCQuality Pr., 1991, p. 127. (Quality and reliability V. 21)

Box, George E.P. and Others Statistics for Experimenters; An Introduction to Design, data Analysis andModel Building. New York, John Wiley, 1978. p. 653. (Wiley Series in Probability andMathematical Statistics).

Boyd, Harper W. and Westfall, Ralph, Marketing Research; Text and Cases. 3rd ed. Illinois: Irwin,1972, p. 813.

British Productivity Council, London, Quality and Reliability; Basic Principles Explained with 16 caseHistories. London, p. 36.

Camp, Robert C. Benchmarking; The Search for Industry Best Practices that Lead to SuperiorPerformance. Wisconsin: ASQC Quality Pr., 1989, p. 399.



Carter (Chuck), C.L.; Jr. Quality Assurance, Quality Control and Inspection Handbook/By C.L. Carter,Jr. and Associates. 3rd. ed., Texas: C.L. Carter, 1979.

Carter, C.L. (Chuck); Jr., The Control and Assurance of Quality Reliability and Safety. Texas: TheAuth., 1978, p. 325.

Crosby, Philip B. Quality Without Tears; the Art of Hasele-Free Management. New York: McGraw-Hill,1984, p. 205.

Croshy, Philip B. Quality is free; The Art of Making Quality Certain. New York: McGraw-Hill, p. 509.

Daniel, Cuthbert, Applications of Statistics to Industrial Experimental. New York: John Wiley, 1976,p. 294.

Das, M.N. and Giri, N.C. Design and Analysis of Experiments�New Delhi: Wiley Eastern, 1979, p. 295.

Deming, W. Edwards, Sample Design in Business Research�New York: John Wiley, 1960, p. 517.

Deming, W. Edwards, Out of the Crisis; Quality Productivity and Competitive Position�Madras:Productivity and Quality Publs., 1992 (Reprint 1993).

Deming, W. Edwards, Quality, Productivity and Competitive Position�Cambridge: MIT Pr., 1982,p. 373.

Drummond, Helga, The Quality Movement; What Total Quality Management is Really All About.London: Kogan Page, 1992, p. 160.

Feigenbaum, A.V. Quality Control; Principles, Practice and Administration�New York: McGraw-Hill,1951, p. 443. (McGraw-Hill Industrial Organization and Management Series).

Ferber, Robert, Statistical Techniques in Market Research. New York: McGraw-Hill, 1949.

Ferber, Robert, Handbook of Marketing Research. New York: McGraw-Hill, 1974.

Flood, Robert Louis, Beyond TQM. Chichester: John Wiley, 1993.

Gilbert, John, How to Eat an Elephant; A Slice by Slice Guide to Total Quality Management. NewDelhi: Affiliated East West Pr., 1992, p. 212.

Goldstone, Leonard, Statistics in the Management of Nursing Services. Kent: Pitman, 1980, p. 271.

Govil, A.K. Reliability Engineering�New Delhi: Tata McGraw-Hill, 1983, p. 330.

Greene, Richard Tabor, Global Quality; A Synthesis of the World�s Best Management Methods.Milwaukee: ASQC Quality Pr., 1993, p. 886.

Greene, Richard Tabor, Global Quality; A Synthesis of the World�s Best Management Methods.Milwaukee: ASQC Quality Pr., 1993, p. 886.

Hammer, Peter L. (Ivanescu) and Rudeanu, Sergiu Boolean Methods in Operations Research andRelated Areas; Preface By Richard Bellman. Berlin: Springer-Verlag, 1968, p. 329. (Econometricsand Operations Research, V. 7).

Harrington, H.J. The Improvement Process; How America�s Leading Companies Improve Quality�NewYork: McGraw-Hill, 1987, p. 239.

Harrington, H. James, Excellence; The IBM Way�Wisconsin: ASQC Quality Pr., 1988, p. 112.

Hemphill, Charles F. Management�s Role in Loss Prevention. New York: American Management Assn.,1976, p. 186.

Hicks, charles R. Fundamental Concepts in the Design of Experiments�3rd Ed., New York: Rinehartand Winston, 1982i, p. 425.

Hutchins, David, Just in Time. England: Gower, 1988, xii, p. 215. Index..

Imai, Masaaki, Kaizen (Ky�Zen); The Key to Japan�s Competitive Success�New York: McGraw-Hill,1986, p. 259.

Indian Standards Institution, New Delhi, Sampling Inspection Tables. New Delhi: The Auth.

BIBLIOGRAPHY 151


International Computers and Tabulators Limited, London, What is O.R.; An Introduction to OperationalResearch Techniques. London: The Auth., 1964, p. 23.

International Academy for Quality, The Best on Quality; Targets, Improvements, Systems. Ed. by JohnD. Hromi., Wisconsin: ASQC Quality Pr., 1996, p. 590.

Ishikawa, Kaoru, What is Total Quality Control? The Japanese Way�New Jersey: Prentice-Hall, 1985,p. 215. �Translated from the Japanese by David J. Lu�.

Japanese Standards Association, Tokyo Guide to Quality Control and Company Standradization�Tokyo: Japanese Standards Association, 1984, p. 121.

Juran, J.M. Course on Management of Quality�4th. ed.�New York: United Nations, 1981.

Juran, J.M. Juran on Planning for Quality�New York: The Free Pr., 1988, p. 341.

Juran, J.M., Juran on Quality by Design; The New Steps for Planning Quality into Goods andServices�New York: The Free Pr., 1992, p. 538.

Juran, J.M., ed. A History of Managing for Quality, The Evolution, Trends, and Future; Directions ofManaging for Quality; A Research Project Sponsored by Juran Foundation. Wisconsin: ASQCQuality Pr., 1995, p. 688.

Juran, J.M. Managerial Breakthrough; A New Concept of the Manager�s Job, New York: McGraw-Hill,1964, p. 396.

Juran, J.M. and Gryan, Frank M. Quality Planning and Analysis from Product Development ThroughUse. 3rd. ed. New York: McGraw-Hill, 1993, p. 634. (McGraw-Hill Series in Industrial Engg. andManagement Science).

Juran, J.M. Juran on Leadership for Quality; and Executive Handbook. New York: The Free Pr., 1989,p. 376.

Kanji, Gopal K. and Asher, Mike, Hundred Methods for Total Quality Management. New Delhi: SAGE,1996, p. 237.

Kanji, Gopal K. and Asher Mike, Total Quality Management Processes; A Systematic Approach.Oxfordshire: Carfax Publs., 1993, p. 132.

Kanti Swarup and Others, Operations Research/By Kanti Swarup, P.K. Gupta and Man Mohan�NewDelhi: Sultan Chand, 1977, p. 728.

Karatsu, Hajime and Ikeda, Toyoki, Mastering the Tools of Q.C.; Learning Through Diagrams andIllustrations. Tokyo: PHP Instt., 1987, p. 243.

Karatsu, Hajime, TQC wisdom of Japan; Managing for Total Quality Control�Cambridge: ProductivityPr., 1980, p. 125.

Kehoe, Dennis F. The Fundamentals of Quality Management. London: Chapman and Hall, 1996, p. 289.

Labovitz, George and Others. Making Quality Work; A Leadership Guide for the Results DrivenManager/By George Labovitz, Yu sang Chang and Victor Rosansky. Essex: Oliver Wight Publns.,1992, p. 198.

Lester, Ronald H. and Others Quality Control for Profit/By Ronald H. Laster, Norbert L. Enrick andHarry E. Mottley. New York: Industrial Pr., 1977.

Logothetis, N. and Wynn, H.P. Quality Through Design; Experimental Design; Off-Line Quality Controland Taguchi�s Contributions�Oxford: Clarendon Pr., 1989, p. 464.

Matsushita, Konosuke, Not for Bread Alone; A Business Ethos, A Management Ethic�Kyoto: PHPInstt., 1984 (Rep. 1987), p. 167.

Mears, Peter and Voehl, Frank, The Executive Guide to Implementing Quality Systems. London: KoganPage, 1995, p. 218.

Milakovich, Michael E. Improving Service Quality; Achieving High Performance in the Public andPrivate Sectors. Florida: St. Lucie Pr., 1995, p. 258.



Miyashita, Kenichi and Russell, David, Keiretsu; Inside the Hidden Japanese Conglomerates. NewYork: McGraw-Hill, 1996, p. 225.

Mizuno, Shigeru; ed. Management for Quality Improvement; The Seven New QC Tools�Cambridge:Productivity Pr., 1988, p. 304.

Mizuno, Shigeru, Company-Wide Total Quality Control�Tokyo: Asian Productivity Organization, 1988,p. 313.

Moder, Joseph J. and Elmaghraby, Salah E., (Editors) Handbook of Operations Research. New York:Van Nostrand, 1978, p. 662.

Montgomery, Douglas C. Design and Analysis of Experiments�New York: John Wiley, 1976, p. 418.

Montgomery, Douglas C. and Peck, Elizabeth A. Introduction to Linear Regression Analysis. New York:John Wiley, 1982, p. 504.

Moss, Marvin A. Applying TQM to Product Design and Development. New York: Marcel Dekker, 1996,p. 412.

Moss, Marvin A. Designing for Minimal Maintenance Expense; The Practical Application of Reliabilityand Maintainability. New York: Marcel Dekker, 1985.

Munro-Faure, Lesley and Others, Achieving Quality Standards; A Step-by Step Guide to BS5750/ISO9000/By Lesley Munro-Faure, Malcolm Munro-Faure and Edward Bones. London: Pitman, 1993, p.210.

Murthy, M.N. Sampling Theory and Methods�Calcutta: Statistical Publs., 1967, p. 684 (StatisticalPublishing Society; Books on Probability and Statistics)

Nelson, Charles, TQM and ISO 9000 for Architects and Designers. New York: McGraw-Hill, 1996,p. 317.

Oakland, John S. and Porter, Leslie J. Cases in Total Quality Management. Oxford: Heinemann, 1994,p. 271. Index, Appendix.

Oakland, John S. Total Quality Management; The Route to Improving Performance�2nd. ed.�Oxford:Heinemann, 1993, p. 463.

Odeh, Robert E. and Fox, Martin, Sample Size Choice; Charts for Experiments with Linear Models�2nd ed.�New York: Marcel Dekker, 1991, p. 194. (Statistics, Textbooks and Monographs; V. 122)

Park, Sung H. Robust Design and Analysis for Quality Engineering. Madras: Chapman and Hall, 1996,p. 329.

Patton, Joseph D., Jr. Maintainability and Maintenance Management�North Carolina: InstrumentSociety of America, 1980, p. 441.

Persico, John, Jr., ed. The TQM Transformation; A Model for Organizational Change�New York:quality Resources, 1992, p. 195.

Plowman, Brian, High Value, Low Cost; How to Create Profitable Customer Delight. London: Pitman,1994, p. 267.

Poirier, Charles C. and Tokarz, Steven J. Avoiding the Pitfalls of Total Quality, Wisconsin: ASQCQuality Pr., 1996, p. 244.

Price, Frank, Right Every Time Using the Deming Approach/By Frank Price., Hants: Gower, 1990, p.182.

Quality Management Library. CCTA, The Govt. Centre for Information Systems. London: HMSO, 1992.1-Overview, p. 15, 2-Quality Management System Implementation/By C.J. Ashby, xv, p. 194, 3-Quality Management System Audit/By C.A. Moore, p. 94, 4-Quality Techniques/By C. Barratt, J.Slater and D. Morgan, ix, p. 75, 5-Quality Training/By C.A. Moore and C.J. Ashby, p. 61.

Rosander, A.C. Applications of Quality Control in the Service Industries�New York: Marcel Dekker,1985, p. 387. Index, Bibl. (Quality and reliability)

BIBLIOGRAPHY 153


Rosander, A.C. The Quest for Quality in Services. Wisconsin: ASQC Quality Pr., 1989, p. 568.

Ross, Joel E. Managing Productivity. Reston: Reston Publns., 1977, p. 191.

Rothery, Brain, ISO 9000 Aldershot. Hampshire: Gower, 1991, p. 182.

Russell, Robert S. and Taylor, Bernard W. Production and Operations Management; Focusing onQuality and Competitiveness. New Jersey: Prentice-Hall, 1995.i, p. 916.

Rust, Roland T. and Others, Return on Quality Measuring the Financial Impact of Your Company�sQuest for Quality/By R.T. Rust, Anthony J. Zahorik and Timothy L. Keiningh. Chicago: IllinoisUniv. Pr., 1994.

Sasaki, Naoto and Hutchins, David, ed. The Japanese Approach to Product Quality; its Applicability tothe West. Collection of Papers Presented to the Past Four Conferences of QC Circle held in Londonsince 1979. Oxford: Pergamon, 1984, p. 134.

Sayre, Don, Inside ISO 14000; The Competitive Advantage of Environmental Management. Florida: St.Lucie Pr., 1996, p. 232.

Seminar on Upper Management and Quality Juran Institute, N.Y. 1988, Upper Nanagement andQuality. 4th ed./By J.M. Juran. New York: Juran Instt., 1982.

Shecter, Edwin S. Managing for Worldclass Quality; A Primer for Executives and Managers. New York:ASQC Quality Pr., 1992.

Shewhart, W.A. Economic Control of Quality of Manufactured Product�New York: Van Nostrand, 1931,p. 501 (Bell Telephone Laboratories Series)

Shores, A. Richard Survival of the Fittest; Total Quality Control and Management Evolution.Wisconsin: ASQC Quality Pr., 1988, p. 302.

Sloan, M. Daniel, Using Designed Experiments to Shrink Health Care Costs. Wisconsin: ASQC, p. 215.

Snee, Ronald D. and Others Experiments in Industry; Design, Analysis and Interpretation of Results/By Ronald D. Snee, Lynne B. Hare and J. Richard Trout Milwaukee. Wisconsin: ASQC Quality Pr.,1985.

Stebbing, Lionel, Quality Assurance; The Route to Efficiency and Competitiveness�3rd. ed.�New York:Ellis Horwood, 1993, p. 336.

Stein, Robert E. The Next Phase of Total Quality Management; TQM II and The Focus on Profitability.New York: Marcell Dekker, 1994.

Stephan, F.F. and McCarthy, P.J. Sampling Opinions; an Analysis of Survey Procedure�New York:John Wiley, 1958, p. 451 (A Wiley Publication of Related Interest to Statisticians)

Swanson Roger C. The Quality Improvement Handbook; Team Guide to Tools and Techniques. London:Kogan Page, 1995, p. 267.

Taguchi, Genichi and Others, Quality Engineering in Production Systems/By Genichi Taugchi, ElsayedA. Elsayed and Thomas Hsiang. New York: McGraw-Hill, 1989, p. 173.

Taguchi, Genichi, Introduction to Quality Engineering; Designing Quality into Products and Processes.Tokyo: Asian Productivity Organization, 1985, p. 191.

Taguchi, Genichi, On-line Quality Control During Production. Tokyo: Japanese Standards Asson., 1981.

Taguchi, Genichi and Wu, Yu-In, Introduction to Off-Line Quality Control. Tokyo: Central JapanQuality Control Asson., p. 115.

Tague, Nency R. The Quality Toolbox. Wisconsin: ASQC Quality Pr., 1996, p. 298.

Thomas, Brian, The Human Dimension of Quality. London: McGraw-Hill, 1995, p. 259.

United Nations. Industrial Development Organisation, Vienna Standardization Based on theProceedings of the International Symposium on Development, Athens, Nov.-Dec. 1967. New York:United Nations, 1969, p. 66. (UNIDO Monographs on Industrialization of Developing Countries)



Wakhlu, Bharat, Total Quality; Excellence Through Organization; Wide Transformation. New Delhi:Wheeler Publns., 1994, p. 272.

Waller, Jenny and Others, The TQM Toolkit; A Guide to Practical Techniques for Total QualityManagement/BY Jenny Waller, Derek Allen and Andrew Burns. London: Kogan Page, 1995.

Walton, Mary, Deming Management at Work. London: Mercury Bks., 1991, p. 252.

Wise, Stephen A. and Fair, Douglas C. Innovative Control Charting; Practical SPC Solutions for Today�sManufacturing Environment. Wisconsin: ASQC Quality Pr., 1998, p. 292.

Zaidi, A. SPC; Concepts, Methodologies and Tools. New Delhi: Prentice-Hall, 1995, p. 353.

Zairi, Mohamed, Total Quality Management for Engineers�New Delhi: Aditya Bks., 1992, p. 245.

Documents

The Seven Magnificent